text,video_id,timestamp,lines,chunk_start,chunk_end
introduction of artificial intelligence and machinelearningby the end of thislessonyou will be abletodefine artificialintelligencedescribe the relationship between artificial intelligence and datasciencedefine machinelearningdescribe the relationship between machinelearningartificial intelligence and datasciencedescribe different machine learningapproachesidentify the applications of machinelearninglet's understand how the field of artificial intelligenceemergedlet's first understand the reason behind the emergence ofa.idata economy is one of the factors behind the emergence ofaiit refers to how much data has grown over the past few years and how much more it can grow in the comingyearswhen you look at thisgraphyou can clearly understand how the volume of data hasgrownyou can see that since2009the data volume has increased by 44timeswith the help of socialwebsitesthe explosion of data has given rise to a new economy and there is a constant battle for ownership of data between companies to derive benefits fromitnow that you know that data has grown at a rapid pace in the past few years and is going to continue togrowlet's understand the need foraias youknowthe increase in data volume has given rise to bigdatawhich helps manage huge amounts ofdata,SSE4M0gcmvE,00:00:10.160,"[{""timestamp"": ""00:00:10.160"", ""start_sec"": 10.16, ""end_sec"": 16.2, ""text"": ""introduction of artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:12.400"", ""start_sec"": 12.4, ""end_sec"": 16.2, ""text"": ""and machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:16.640"", ""start_sec"": 16.64, ""end_sec"": 21.68, ""text"": ""by the end of this lesson you will be"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:18.480"", ""start_sec"": 18.48, ""end_sec"": 23.519, ""text"": ""able to define artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:21.680"", ""start_sec"": 21.68, ""end_sec"": 26.88, ""text"": ""describe the relationship between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:23.519"", ""start_sec"": 23.519, ""end_sec"": 29.439, ""text"": ""artificial intelligence and data science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:26.880"", ""start_sec"": 26.88, ""end_sec"": 31.119, ""text"": ""define machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:29.439"", ""start_sec"": 29.439, ""end_sec"": 33.44, ""text"": ""describe the relationship between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:31.119"", ""start_sec"": 31.119, ""end_sec"": 35.76, ""text"": ""machine learning artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:33.440"", ""start_sec"": 33.44, ""end_sec"": 37.76, ""text"": ""and data science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:35.760"", ""start_sec"": 35.76, ""end_sec"": 38.96, ""text"": ""describe different machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:37.760"", ""start_sec"": 37.76, ""end_sec"": 41.2, ""text"": ""approaches"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:38.960"", ""start_sec"": 38.96, ""end_sec"": 43.36, ""text"": ""identify the applications of machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:41.200"", ""start_sec"": 41.2, ""end_sec"": 45.28, ""text"": ""learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:43.360"", ""start_sec"": 43.36, ""end_sec"": 48.48, ""text"": ""let's understand how the field of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:45.280"", ""start_sec"": 45.28, ""end_sec"": 50.64, ""text"": ""artificial intelligence emerged"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:48.480"", ""start_sec"": 48.48, ""end_sec"": 52.64, ""text"": ""let's first understand the reason behind"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:50.640"", ""start_sec"": 50.64, ""end_sec"": 54.719, ""text"": ""the emergence of a.i"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:52.640"", ""start_sec"": 52.64, ""end_sec"": 57.039, ""text"": ""data economy is one of the factors"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:54.719"", ""start_sec"": 54.719, ""end_sec"": 59.359, ""text"": ""behind the emergence of ai"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:57.039"", ""start_sec"": 57.039, ""end_sec"": 61.92, ""text"": ""it refers to how much data has grown"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:00:59.359"", ""start_sec"": 59.359, ""end_sec"": 64.239, ""text"": ""over the past few years and how much"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:01.920"", ""start_sec"": 61.92, ""end_sec"": 66.159, ""text"": ""more it can grow in the coming years"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:04.239"", ""start_sec"": 64.239, ""end_sec"": 68.32, ""text"": ""when you look at this graph you can"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:06.159"", ""start_sec"": 66.159, ""end_sec"": 70.0, ""text"": ""clearly understand how the volume of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:08.320"", ""start_sec"": 68.32, ""end_sec"": 72.72, ""text"": ""data has grown"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:10.000"", ""start_sec"": 70.0, ""end_sec"": 75.84, ""text"": ""you can see that since 2009 the data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:12.720"", ""start_sec"": 72.72, ""end_sec"": 78.0, ""text"": ""volume has increased by 44 times with"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:15.840"", ""start_sec"": 75.84, ""end_sec"": 80.56, ""text"": ""the help of social websites"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:18.000"", ""start_sec"": 78.0, ""end_sec"": 83.119, ""text"": ""the explosion of data has given rise to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:20.560"", ""start_sec"": 80.56, ""end_sec"": 85.36, ""text"": ""a new economy and there is a constant"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:23.119"", ""start_sec"": 83.119, ""end_sec"": 89.03999999999999, ""text"": ""battle for ownership of data between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:25.360"", ""start_sec"": 85.36, ""end_sec"": 91.119, ""text"": ""companies to derive benefits from it"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:29.040"", ""start_sec"": 89.03999999999999, ""end_sec"": 93.68, ""text"": ""now that you know that data has grown at"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:31.119"", ""start_sec"": 91.119, ""end_sec"": 96.0, ""text"": ""a rapid pace in the past few years and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:33.680"", ""start_sec"": 93.68, ""end_sec"": 98.56, ""text"": ""is going to continue to grow"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:36.000"", ""start_sec"": 96.0, ""end_sec"": 100.96000000000001, ""text"": ""let's understand the need for ai"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:38.560"", ""start_sec"": 98.56, ""end_sec"": 104.0, ""text"": ""as you know the increase in data volume"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:40.960"", ""start_sec"": 100.96000000000001, ""end_sec"": 106.32, ""text"": ""has given rise to big data which helps"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:44.000"", ""start_sec"": 104.0, ""end_sec"": 109.52000000000001, ""text"": ""manage huge amounts of data"", ""video_id"": ""SSE4M0gcmvE""}]",10.16,109.52000000000001
data science helps analyze thatdataso the science associated with data is going toward a new paradigm where one can teach machines to learn from data and drive a variety of usefulinsightsgiving rise to artificialintelligencenow you mayaskwhat is artificialintelligenceartificial intelligence refers to the intelligence displayed by machines that simulates human and animalintelligenceit involves intelligenceagentsthe autonomous entities that perceive their environment and take actions that maximize their chances of success at a givengoalartificial intelligence is a technique that enables computers to mimic human intelligence usinglogicit is a program that cansensereason andactlet's look at some of the areas where artificial intelligence isusedartificial intelligence is redefining industries by providing greater personalization to users and automatingprocessesone example of artificial intelligence in practice is self-drivingcarsself-driving cars are computer controlled cars that drivethemselvesin thesecarshuman drivers are never required to take control to safely operate thevehiclethese cars are also known as autonomous or driverlesscarslet's see how apple usesaiiphone users can experience the power ofsirithevoice,SSE4M0gcmvE,00:01:46.320,"[{""timestamp"": ""00:01:46.320"", ""start_sec"": 106.32, ""end_sec"": 112.0, ""text"": ""data science helps analyze that data so"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:49.520"", ""start_sec"": 109.52000000000001, ""end_sec"": 114.64, ""text"": ""the science associated with data is"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:52.000"", ""start_sec"": 112.0, ""end_sec"": 116.72, ""text"": ""going toward a new paradigm"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:54.640"", ""start_sec"": 114.64, ""end_sec"": 119.6, ""text"": ""where one can teach machines to learn"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:56.720"", ""start_sec"": 116.72, ""end_sec"": 121.84, ""text"": ""from data and drive a variety of useful"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:01:59.600"", ""start_sec"": 119.6, ""end_sec"": 124.159, ""text"": ""insights giving rise to artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:01.840"", ""start_sec"": 121.84, ""end_sec"": 126.84, ""text"": ""intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:04.159"", ""start_sec"": 124.159, ""end_sec"": 129.36, ""text"": ""now you may ask what is artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:06.840"", ""start_sec"": 126.84, ""end_sec"": 131.68, ""text"": ""intelligence artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:09.360"", ""start_sec"": 129.36, ""end_sec"": 134.64, ""text"": ""refers to the intelligence displayed by"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:11.680"", ""start_sec"": 131.68, ""end_sec"": 135.84, ""text"": ""machines that simulates human and animal"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:14.640"", ""start_sec"": 134.64, ""end_sec"": 138.319, ""text"": ""intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:15.840"", ""start_sec"": 135.84, ""end_sec"": 140.4, ""text"": ""it involves intelligence agents"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:18.319"", ""start_sec"": 138.319, ""end_sec"": 143.12, ""text"": ""the autonomous entities that perceive"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:20.400"", ""start_sec"": 140.4, ""end_sec"": 145.36, ""text"": ""their environment and take actions that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:23.120"", ""start_sec"": 143.12, ""end_sec"": 146.8, ""text"": ""maximize their chances of success at a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:25.360"", ""start_sec"": 145.36, ""end_sec"": 148.959, ""text"": ""given goal"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:26.800"", ""start_sec"": 146.8, ""end_sec"": 151.28, ""text"": ""artificial intelligence is a technique"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:28.959"", ""start_sec"": 148.959, ""end_sec"": 153.84, ""text"": ""that enables computers to mimic human"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:31.280"", ""start_sec"": 151.28, ""end_sec"": 156.56, ""text"": ""intelligence using logic"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:33.840"", ""start_sec"": 153.84, ""end_sec"": 158.4, ""text"": ""it is a program that can sense reason"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:36.560"", ""start_sec"": 156.56, ""end_sec"": 160.31900000000002, ""text"": ""and act"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:38.400"", ""start_sec"": 158.4, ""end_sec"": 163.2, ""text"": ""let's look at some of the areas where"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:40.319"", ""start_sec"": 160.31900000000002, ""end_sec"": 165.68, ""text"": ""artificial intelligence is used"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:43.200"", ""start_sec"": 163.2, ""end_sec"": 167.519, ""text"": ""artificial intelligence is redefining"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:45.680"", ""start_sec"": 165.68, ""end_sec"": 170.48, ""text"": ""industries by providing greater"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:47.519"", ""start_sec"": 167.519, ""end_sec"": 171.76, ""text"": ""personalization to users and automating"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:50.480"", ""start_sec"": 170.48, ""end_sec"": 174.0, ""text"": ""processes"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:51.760"", ""start_sec"": 171.76, ""end_sec"": 177.519, ""text"": ""one example of artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:54.000"", ""start_sec"": 174.0, ""end_sec"": 179.36, ""text"": ""in practice is self-driving cars"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:57.519"", ""start_sec"": 177.519, ""end_sec"": 182.4, ""text"": ""self-driving cars are computer"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:02:59.360"", ""start_sec"": 179.36, ""end_sec"": 184.64, ""text"": ""controlled cars that drive themselves"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:02.400"", ""start_sec"": 182.4, ""end_sec"": 186.8, ""text"": ""in these cars human drivers are never"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:04.640"", ""start_sec"": 184.64, ""end_sec"": 188.48, ""text"": ""required to take control to safely"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:06.800"", ""start_sec"": 186.8, ""end_sec"": 191.12, ""text"": ""operate the vehicle"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:08.480"", ""start_sec"": 188.48, ""end_sec"": 193.599, ""text"": ""these cars are also known as autonomous"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:11.120"", ""start_sec"": 191.12, ""end_sec"": 196.56, ""text"": ""or driverless cars"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:13.599"", ""start_sec"": 193.599, ""end_sec"": 199.2, ""text"": ""let's see how apple uses ai"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:16.560"", ""start_sec"": 196.56, ""end_sec"": 200.879, ""text"": ""iphone users can experience the power of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:19.200"", ""start_sec"": 199.2, ""end_sec"": 203.04, ""text"": ""siri the voice"", ""video_id"": ""SSE4M0gcmvE""}]",106.32,203.04
it simplifies navigating through your iphone as it listens to your voice commands to performtasksforinstanceyou can ask siri to call your friend or to playmusicsiri is fun and is extremely convenient touseanother example is google'salphagowhich is a computer program that plays the board gamegoit is the first computer program to defeat a world champion at the ancient chinese game ofgoamazon echo is anotherproductit's a home control chatbot device that responds tohumansaccording to what they aresayingit responds by playingmusicmovies andmoreif you've got compatible smart homedevicesyou can tell echo to dim the lights or turn appliances on oroffyou can use ai andchessand here is an example of a concierge robot from ibm called ibmwatsonthe ibm watson ai has typically been in the headlines for composingmusicplaying chess and even cookingfoodlet's move ahead and look at some sci-fi movies with the concept of artificialintelligencethe films featuring ai reflect the ever-changing spectrum of our emotions regarding the machines we havecreatedhumans are fascinated by the concept of artificial intelligence and this is reflected in the wide range of movies onai recommendationssystems are used by a lot of e-commercecompanieslet's see how theyworkamazon collects data from users and,SSE4M0gcmvE,00:03:20.879,"[{""timestamp"": ""00:03:20.879"", ""start_sec"": 200.879, ""end_sec"": 204.8, ""text"": ""it simplifies navigating through your"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:23.040"", ""start_sec"": 203.04, ""end_sec"": 207.44, ""text"": ""iphone as it listens to your voice"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:24.800"", ""start_sec"": 204.8, ""end_sec"": 209.599, ""text"": ""commands to perform tasks"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:27.440"", ""start_sec"": 207.44, ""end_sec"": 212.879, ""text"": ""for instance you can ask siri to call"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:29.599"", ""start_sec"": 209.599, ""end_sec"": 216.07999999999998, ""text"": ""your friend or to play music siri is fun"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:32.879"", ""start_sec"": 212.879, ""end_sec"": 218.48, ""text"": ""and is extremely convenient to use"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:36.080"", ""start_sec"": 216.07999999999998, ""end_sec"": 220.56, ""text"": ""another example is google's alphago"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:38.480"", ""start_sec"": 218.48, ""end_sec"": 222.48, ""text"": ""which is a computer program that plays"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:40.560"", ""start_sec"": 220.56, ""end_sec"": 224.879, ""text"": ""the board game go"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:42.480"", ""start_sec"": 222.48, ""end_sec"": 227.599, ""text"": ""it is the first computer program to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:44.879"", ""start_sec"": 224.879, ""end_sec"": 229.84, ""text"": ""defeat a world champion at the ancient"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:47.599"", ""start_sec"": 227.599, ""end_sec"": 232.56, ""text"": ""chinese game of go"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:49.840"", ""start_sec"": 229.84, ""end_sec"": 234.64, ""text"": ""amazon echo is another product it's a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:52.560"", ""start_sec"": 232.56, ""end_sec"": 237.12, ""text"": ""home control chatbot device that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:54.640"", ""start_sec"": 234.64, ""end_sec"": 239.68, ""text"": ""responds to humans according to what"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:57.120"", ""start_sec"": 237.12, ""end_sec"": 242.08, ""text"": ""they are saying it responds by playing"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:03:59.680"", ""start_sec"": 239.68, ""end_sec"": 244.08, ""text"": ""music movies and more"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:02.080"", ""start_sec"": 242.08, ""end_sec"": 246.239, ""text"": ""if you've got compatible smart home"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:04.080"", ""start_sec"": 244.08, ""end_sec"": 249.12, ""text"": ""devices you can tell echo to dim the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:06.239"", ""start_sec"": 246.239, ""end_sec"": 251.439, ""text"": ""lights or turn appliances on or off you"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:09.120"", ""start_sec"": 249.12, ""end_sec"": 254.879, ""text"": ""can use ai and chess and here is an"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:11.439"", ""start_sec"": 251.439, ""end_sec"": 256.959, ""text"": ""example of a concierge robot from ibm"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:14.879"", ""start_sec"": 254.879, ""end_sec"": 260.079, ""text"": ""called ibm watson"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:16.959"", ""start_sec"": 256.959, ""end_sec"": 262.32, ""text"": ""the ibm watson ai has typically been in"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:20.079"", ""start_sec"": 260.079, ""end_sec"": 266.0, ""text"": ""the headlines for composing music"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:22.320"", ""start_sec"": 262.32, ""end_sec"": 268.32, ""text"": ""playing chess and even cooking food"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:26.000"", ""start_sec"": 266.0, ""end_sec"": 270.56, ""text"": ""let's move ahead and look at some sci-fi"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:28.320"", ""start_sec"": 268.32, ""end_sec"": 272.0, ""text"": ""movies with the concept of artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:30.560"", ""start_sec"": 270.56, ""end_sec"": 274.32, ""text"": ""intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:32.000"", ""start_sec"": 272.0, ""end_sec"": 276.8, ""text"": ""the films featuring ai reflect the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:34.320"", ""start_sec"": 274.32, ""end_sec"": 279.52, ""text"": ""ever-changing spectrum of our emotions"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:36.800"", ""start_sec"": 276.8, ""end_sec"": 281.84000000000003, ""text"": ""regarding the machines we have created"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:39.520"", ""start_sec"": 279.52, ""end_sec"": 284.08, ""text"": ""humans are fascinated by the concept of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:41.840"", ""start_sec"": 281.84000000000003, ""end_sec"": 286.639, ""text"": ""artificial intelligence and this is"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:44.080"", ""start_sec"": 284.08, ""end_sec"": 288.24, ""text"": ""reflected in the wide range of movies on"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:46.639"", ""start_sec"": 286.639, ""end_sec"": 290.8, ""text"": ""ai"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:48.240"", ""start_sec"": 288.24, ""end_sec"": 293.36, ""text"": ""recommendations systems are used by a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:50.800"", ""start_sec"": 290.8, ""end_sec"": 296.15999999999997, ""text"": ""lot of e-commerce companies let's see"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:53.360"", ""start_sec"": 293.36, ""end_sec"": 298.56, ""text"": ""how they work"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:04:56.160"", ""start_sec"": 296.15999999999997, ""end_sec"": 300.8, ""text"": ""amazon collects data from users and"", ""video_id"": ""SSE4M0gcmvE""}]",200.879,300.8
recommends the best product according to the user's buying or shoppingpatternforexamplewhen you search for a specific product in the amazon store and add it to yourcartamazon recommends some relevant products based on your past shopping and searchingpatternso before you buy the selectedproductyou get recommendations based on your interest and there is a possibility that you may also buy the relevant product with a selectedproductifnotyou have the chance to compare the selected product with the recommendedproductsnow let's move ahead and understand the relationship between artificialintelligencemachine learning and datascienceeven though the terms artificialintelligence aimachine learning and data science fall in the same domain and are connected to eachotherthey have their specific applications andmeaninglet's try to understand a little about each of thesetermsartificial intelligence systems mimic or replicate humanintelligencemachine learning provides systems the ability to automatically learn and improve from the experiences without being explicitlyprogrammeddata science is an umbrella term that encompasses dataanalyticsdataminingmachinelearningartificial intelligence and several other relateddisciplineslet's look at the flow diagram and try to understand the relationship betweenaimachine learning and datascience,SSE4M0gcmvE,00:04:58.560,"[{""timestamp"": ""00:04:58.560"", ""start_sec"": 298.56, ""end_sec"": 303.759, ""text"": ""recommends the best product according to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:00.800"", ""start_sec"": 300.8, ""end_sec"": 305.52, ""text"": ""the user's buying or shopping pattern"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:03.759"", ""start_sec"": 303.759, ""end_sec"": 308.72, ""text"": ""for example when you search for a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:05.520"", ""start_sec"": 305.52, ""end_sec"": 310.4, ""text"": ""specific product in the amazon store and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:08.720"", ""start_sec"": 308.72, ""end_sec"": 312.8, ""text"": ""add it to your cart"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:10.400"", ""start_sec"": 310.4, ""end_sec"": 314.479, ""text"": ""amazon recommends some relevant products"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:12.800"", ""start_sec"": 312.8, ""end_sec"": 316.16, ""text"": ""based on your past shopping and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:14.479"", ""start_sec"": 314.479, ""end_sec"": 318.479, ""text"": ""searching pattern"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:16.160"", ""start_sec"": 316.16, ""end_sec"": 320.4, ""text"": ""so before you buy the selected product"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:18.479"", ""start_sec"": 318.479, ""end_sec"": 322.96, ""text"": ""you get recommendations based on your"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:20.400"", ""start_sec"": 320.4, ""end_sec"": 325.199, ""text"": ""interest and there is a possibility that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:22.960"", ""start_sec"": 322.96, ""end_sec"": 328.24, ""text"": ""you may also buy the relevant product"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:25.199"", ""start_sec"": 325.199, ""end_sec"": 330.0, ""text"": ""with a selected product if not you have"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:28.240"", ""start_sec"": 328.24, ""end_sec"": 333.28, ""text"": ""the chance to compare the selected"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:30.000"", ""start_sec"": 330.0, ""end_sec"": 335.36, ""text"": ""product with the recommended products"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:33.280"", ""start_sec"": 333.28, ""end_sec"": 337.039, ""text"": ""now let's move ahead and understand the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:35.360"", ""start_sec"": 335.36, ""end_sec"": 339.84000000000003, ""text"": ""relationship between artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:37.039"", ""start_sec"": 337.039, ""end_sec"": 342.0, ""text"": ""intelligence machine learning and data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:39.840"", ""start_sec"": 339.84000000000003, ""end_sec"": 343.68, ""text"": ""science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:42.000"", ""start_sec"": 342.0, ""end_sec"": 346.56, ""text"": ""even though the terms artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:43.680"", ""start_sec"": 343.68, ""end_sec"": 349.52, ""text"": ""intelligence ai machine learning and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:46.560"", ""start_sec"": 346.56, ""end_sec"": 351.36, ""text"": ""data science fall in the same domain and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:49.520"", ""start_sec"": 349.52, ""end_sec"": 354.15999999999997, ""text"": ""are connected to each other they have"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:51.360"", ""start_sec"": 351.36, ""end_sec"": 356.32, ""text"": ""their specific applications and meaning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:54.160"", ""start_sec"": 354.15999999999997, ""end_sec"": 358.0, ""text"": ""let's try to understand a little about"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:56.320"", ""start_sec"": 356.32, ""end_sec"": 360.479, ""text"": ""each of these terms"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:05:58.000"", ""start_sec"": 358.0, ""end_sec"": 362.8, ""text"": ""artificial intelligence systems mimic or"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:00.479"", ""start_sec"": 360.479, ""end_sec"": 364.88, ""text"": ""replicate human intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:02.800"", ""start_sec"": 362.8, ""end_sec"": 367.44, ""text"": ""machine learning provides systems the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:04.880"", ""start_sec"": 364.88, ""end_sec"": 369.44, ""text"": ""ability to automatically learn and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:07.440"", ""start_sec"": 367.44, ""end_sec"": 372.0, ""text"": ""improve from the experiences without"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:09.440"", ""start_sec"": 369.44, ""end_sec"": 374.16, ""text"": ""being explicitly programmed"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:12.000"", ""start_sec"": 372.0, ""end_sec"": 377.36, ""text"": ""data science is an umbrella term that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:14.160"", ""start_sec"": 374.16, ""end_sec"": 380.24, ""text"": ""encompasses data analytics data mining"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:17.360"", ""start_sec"": 377.36, ""end_sec"": 382.88, ""text"": ""machine learning artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:20.240"", ""start_sec"": 380.24, ""end_sec"": 385.12, ""text"": ""and several other related disciplines"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:22.880"", ""start_sec"": 382.88, ""end_sec"": 386.96, ""text"": ""let's look at the flow diagram and try"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:25.120"", ""start_sec"": 385.12, ""end_sec"": 387.919, ""text"": ""to understand the relationship between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:26.960"", ""start_sec"": 386.96, ""end_sec"": 390.8, ""text"": ""ai"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:27.919"", ""start_sec"": 387.919, ""end_sec"": 393.68, ""text"": ""machine learning and data science"", ""video_id"": ""SSE4M0gcmvE""}]",298.56,393.68
interestinglyml is also an element of artificialintelligenceso the first step is data gathering and datatransformationthis step basically comes under datasciencedata transformation is the process of converting data from one format or structure into another format orstructuredata transformation is important to activities such as data management and dataintegrationafter gatheringdatawe would want to use the data to make predictions and deriveinsightsin order to get predictions out of the datasetwe use machine learning techniques such as supervised learning or unsupervisedlearningon an overviewlevelsupervised and unsupervised learning are the machine learning techniques used to extract predictions from a given datasetnow you must be thinking where deep learning comes into thepicturedeep learning is a subfield of machine learning involved withalgorithmsit uses artificial neural networks which are modeled on the structure and performance of neurons in the humanbraindeep learning is most effective when there isn't a clear structure to the data that you can just exploit and build featuresaroundnow the next step in the flow diagram is to get insights from predictions beingmade,SSE4M0gcmvE,00:06:30.800,"[{""timestamp"": ""00:06:30.800"", ""start_sec"": 390.8, ""end_sec"": 395.68, ""text"": ""interestingly ml is also an element of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:33.680"", ""start_sec"": 393.68, ""end_sec"": 398.15999999999997, ""text"": ""artificial intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:35.680"", ""start_sec"": 395.68, ""end_sec"": 400.0, ""text"": ""so the first step is data gathering and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:38.160"", ""start_sec"": 398.15999999999997, ""end_sec"": 402.319, ""text"": ""data transformation"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:40.000"", ""start_sec"": 400.0, ""end_sec"": 403.52, ""text"": ""this step basically comes under data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:42.319"", ""start_sec"": 402.319, ""end_sec"": 405.6, ""text"": ""science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:43.520"", ""start_sec"": 403.52, ""end_sec"": 407.759, ""text"": ""data transformation is the process of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:45.600"", ""start_sec"": 405.6, ""end_sec"": 410.08, ""text"": ""converting data from one format or"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:47.759"", ""start_sec"": 407.759, ""end_sec"": 411.199, ""text"": ""structure into another format or"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:50.080"", ""start_sec"": 410.08, ""end_sec"": 413.199, ""text"": ""structure"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:51.199"", ""start_sec"": 411.199, ""end_sec"": 415.759, ""text"": ""data transformation is important to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:53.199"", ""start_sec"": 413.199, ""end_sec"": 417.759, ""text"": ""activities such as data management and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:55.759"", ""start_sec"": 415.759, ""end_sec"": 419.759, ""text"": ""data integration"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:57.759"", ""start_sec"": 417.759, ""end_sec"": 422.0, ""text"": ""after gathering data we would want to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:06:59.759"", ""start_sec"": 419.759, ""end_sec"": 424.319, ""text"": ""use the data to make predictions and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:02.000"", ""start_sec"": 422.0, ""end_sec"": 426.88, ""text"": ""derive insights in order to get"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:04.319"", ""start_sec"": 424.319, ""end_sec"": 428.639, ""text"": ""predictions out of the data set we use"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:06.880"", ""start_sec"": 426.88, ""end_sec"": 431.599, ""text"": ""machine learning techniques such as"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:08.639"", ""start_sec"": 428.639, ""end_sec"": 434.8, ""text"": ""supervised learning or unsupervised"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:11.599"", ""start_sec"": 431.599, ""end_sec"": 436.639, ""text"": ""learning on an overview level supervised"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:14.800"", ""start_sec"": 434.8, ""end_sec"": 438.96, ""text"": ""and unsupervised learning are the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:16.639"", ""start_sec"": 436.639, ""end_sec"": 441.199, ""text"": ""machine learning techniques used to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:18.960"", ""start_sec"": 438.96, ""end_sec"": 442.4, ""text"": ""extract predictions from a given data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:21.199"", ""start_sec"": 441.199, ""end_sec"": 444.88, ""text"": ""set"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:22.400"", ""start_sec"": 442.4, ""end_sec"": 447.199, ""text"": ""now you must be thinking where deep"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:24.880"", ""start_sec"": 444.88, ""end_sec"": 449.84, ""text"": ""learning comes into the picture"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:27.199"", ""start_sec"": 447.199, ""end_sec"": 452.8, ""text"": ""deep learning is a subfield of machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:29.840"", ""start_sec"": 449.84, ""end_sec"": 455.68, ""text"": ""learning involved with algorithms"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:32.800"", ""start_sec"": 452.8, ""end_sec"": 457.199, ""text"": ""it uses artificial neural networks which"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:35.680"", ""start_sec"": 455.68, ""end_sec"": 459.199, ""text"": ""are modeled on the structure and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:37.199"", ""start_sec"": 457.199, ""end_sec"": 460.56, ""text"": ""performance of neurons in the human"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:39.199"", ""start_sec"": 459.199, ""end_sec"": 462.88, ""text"": ""brain"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:40.560"", ""start_sec"": 460.56, ""end_sec"": 465.199, ""text"": ""deep learning is most effective when"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:42.880"", ""start_sec"": 462.88, ""end_sec"": 466.15999999999997, ""text"": ""there isn't a clear structure to the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:45.199"", ""start_sec"": 465.199, ""end_sec"": 468.15999999999997, ""text"": ""data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:46.160"", ""start_sec"": 466.15999999999997, ""end_sec"": 469.68, ""text"": ""that you can just exploit and build"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:48.160"", ""start_sec"": 468.15999999999997, ""end_sec"": 472.24, ""text"": ""features around"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:49.680"", ""start_sec"": 469.68, ""end_sec"": 474.24, ""text"": ""now the next step in the flow diagram is"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:52.240"", ""start_sec"": 472.24, ""end_sec"": 475.199, ""text"": ""to get insights from predictions being"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:54.240"", ""start_sec"": 474.24, ""end_sec"": 477.68, ""text"": ""made"", ""video_id"": ""SSE4M0gcmvE""}]",390.8,477.68
in order to dosoyou need to use dataanalysiswhich actually is the process under datasciencenowwhen you are done with all oftheseyou must want your data to perform someactionsthis is where ai comes into thepictureartificial intelligence combines predictions and insights to perform actions based on the human decision and automateddecisionnow let's move ahead and understand the relationship between artificialintelligencemachine learning and datasciencelet's look at the relationship between artificial intelligence and machinelearningartificial intelligence is the engineering of making intelligent machines andprogramsmachine learning provides systems the ability to learn from past experiences without being explicitlyprogrammedmachine learning allows machines to gainintelligencethereby enabling artificialintelligencelet's now understand the relationship between machine learning and datasciencedata science and machine learning go hand inhanddata science helps evaluate data for machine learningalgorithmsdata science covers the whole spectrum of dataprocessingwhile machine learning has the algorithmic or statisticalaspectsdata science is the use of statistical methods to find patterns in thedata,SSE4M0gcmvE,00:07:55.199,"[{""timestamp"": ""00:07:55.199"", ""start_sec"": 475.199, ""end_sec"": 480.319, ""text"": ""in order to do so you need to use data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:07:57.680"", ""start_sec"": 477.68, ""end_sec"": 482.4, ""text"": ""analysis which actually is the process"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:00.319"", ""start_sec"": 480.319, ""end_sec"": 484.08, ""text"": ""under data science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:02.400"", ""start_sec"": 482.4, ""end_sec"": 486.24, ""text"": ""now when you are done with all of these"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:04.080"", ""start_sec"": 484.08, ""end_sec"": 487.28, ""text"": ""you must want your data to perform some"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:06.240"", ""start_sec"": 486.24, ""end_sec"": 490.16, ""text"": ""actions"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:07.280"", ""start_sec"": 487.28, ""end_sec"": 492.08, ""text"": ""this is where ai comes into the picture"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:10.160"", ""start_sec"": 490.16, ""end_sec"": 494.479, ""text"": ""artificial intelligence combines"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:12.080"", ""start_sec"": 492.08, ""end_sec"": 497.039, ""text"": ""predictions and insights to perform"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:14.479"", ""start_sec"": 494.479, ""end_sec"": 499.68, ""text"": ""actions based on the human decision and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:17.039"", ""start_sec"": 497.039, ""end_sec"": 501.759, ""text"": ""automated decision"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:19.680"", ""start_sec"": 499.68, ""end_sec"": 503.44, ""text"": ""now let's move ahead and understand the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:21.759"", ""start_sec"": 501.759, ""end_sec"": 506.24, ""text"": ""relationship between artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:23.440"", ""start_sec"": 503.44, ""end_sec"": 508.16, ""text"": ""intelligence machine learning and data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:26.240"", ""start_sec"": 506.24, ""end_sec"": 510.0, ""text"": ""science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:28.160"", ""start_sec"": 508.16, ""end_sec"": 512.08, ""text"": ""let's look at the relationship between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:30.000"", ""start_sec"": 510.0, ""end_sec"": 513.2, ""text"": ""artificial intelligence and machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:32.080"", ""start_sec"": 512.08, ""end_sec"": 515.12, ""text"": ""learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:33.200"", ""start_sec"": 513.2, ""end_sec"": 516.9590000000001, ""text"": ""artificial intelligence is the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:35.120"", ""start_sec"": 515.12, ""end_sec"": 519.039, ""text"": ""engineering of making intelligent"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:36.959"", ""start_sec"": 516.9590000000001, ""end_sec"": 521.279, ""text"": ""machines and programs"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:39.039"", ""start_sec"": 519.039, ""end_sec"": 524.08, ""text"": ""machine learning provides systems the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:41.279"", ""start_sec"": 521.279, ""end_sec"": 526.9590000000001, ""text"": ""ability to learn from past experiences"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:44.080"", ""start_sec"": 524.08, ""end_sec"": 529.36, ""text"": ""without being explicitly programmed"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:46.959"", ""start_sec"": 526.9590000000001, ""end_sec"": 531.8389999999999, ""text"": ""machine learning allows machines to gain"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:49.360"", ""start_sec"": 529.36, ""end_sec"": 534.8389999999999, ""text"": ""intelligence thereby enabling artificial"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:51.839"", ""start_sec"": 531.8389999999999, ""end_sec"": 534.8389999999999, ""text"": ""intelligence"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:54.959"", ""start_sec"": 534.9590000000001, ""end_sec"": 538.88, ""text"": ""let's now understand the relationship"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:56.800"", ""start_sec"": 536.8, ""end_sec"": 540.0, ""text"": ""between machine learning and data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:08:58.880"", ""start_sec"": 538.88, ""end_sec"": 542.16, ""text"": ""science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:00.000"", ""start_sec"": 540.0, ""end_sec"": 543.6, ""text"": ""data science and machine learning go"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:02.160"", ""start_sec"": 542.16, ""end_sec"": 546.32, ""text"": ""hand in hand"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:03.600"", ""start_sec"": 543.6, ""end_sec"": 548.64, ""text"": ""data science helps evaluate data for"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:06.320"", ""start_sec"": 546.32, ""end_sec"": 550.64, ""text"": ""machine learning algorithms"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:08.640"", ""start_sec"": 548.64, ""end_sec"": 553.04, ""text"": ""data science covers the whole spectrum"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:10.640"", ""start_sec"": 550.64, ""end_sec"": 554.88, ""text"": ""of data processing while machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:13.040"", ""start_sec"": 553.04, ""end_sec"": 558.24, ""text"": ""learning has the algorithmic or"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:14.880"", ""start_sec"": 554.88, ""end_sec"": 558.24, ""text"": ""statistical aspects"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:18.640"", ""start_sec"": 558.64, ""end_sec"": 563.8389999999999, ""text"": ""data science is the use of statistical"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:21.040"", ""start_sec"": 561.04, ""end_sec"": 566.08, ""text"": ""methods to find patterns in the data"", ""video_id"": ""SSE4M0gcmvE""}]",475.199,566.08
statistical machine learning uses the same techniques as datasciencedata science includes various techniques like statisticalmodelingvisualization and patternrecognitionmachine learning focuses on developing algorithms from the data provided by makingpredictionsso what is machinelearningmachine learning is the capability of an artificial intelligence system to learn by extracting patterns fromdatait usually deliversquickermore accurate results to help you spot profitable opportunities or dangerousrisksnow you must be curious to understand the features of machinelearningmachine learning uses the data to detect patterns in a data set and adjust program actionsaccordinglypattern detection can be defined as the classification of data based on knowledge already gained or on statistical information extracted from thepatternsit focuses on the development of computer programs that can teach themselves to grow and change when exposed to new data by using a method called reinforcementlearningit uses external feedback to teach the system to change its internal workings in order to guess better nexttimeit enables computers to find hidden insights using iterative algorithms without being explicitlyprogrammedmachine learning uses algorithms that learn from previous data to help produce reliable and repeatabledecisionsit automates analytical modelbuilding,SSE4M0gcmvE,00:09:23.839,"[{""timestamp"": ""00:09:23.839"", ""start_sec"": 563.8389999999999, ""end_sec"": 568.72, ""text"": ""statistical machine learning uses the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:26.080"", ""start_sec"": 566.08, ""end_sec"": 571.12, ""text"": ""same techniques as data science"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:28.720"", ""start_sec"": 568.72, ""end_sec"": 574.0, ""text"": ""data science includes various techniques"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:31.120"", ""start_sec"": 571.12, ""end_sec"": 577.04, ""text"": ""like statistical modeling visualization"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:34.000"", ""start_sec"": 574.0, ""end_sec"": 579.44, ""text"": ""and pattern recognition machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:37.040"", ""start_sec"": 577.04, ""end_sec"": 584.08, ""text"": ""focuses on developing algorithms from"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:39.440"", ""start_sec"": 579.44, ""end_sec"": 587.68, ""text"": ""the data provided by making predictions"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:44.080"", ""start_sec"": 584.08, ""end_sec"": 589.92, ""text"": ""so what is machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:47.680"", ""start_sec"": 587.68, ""end_sec"": 592.56, ""text"": ""machine learning is the capability of an"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:49.920"", ""start_sec"": 589.92, ""end_sec"": 595.36, ""text"": ""artificial intelligence system to learn"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:52.560"", ""start_sec"": 592.56, ""end_sec"": 597.6, ""text"": ""by extracting patterns from data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:55.360"", ""start_sec"": 595.36, ""end_sec"": 599.6, ""text"": ""it usually delivers quicker more"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:57.600"", ""start_sec"": 597.6, ""end_sec"": 602.48, ""text"": ""accurate results to help you spot"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:09:59.600"", ""start_sec"": 599.6, ""end_sec"": 604.399, ""text"": ""profitable opportunities or dangerous"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:02.480"", ""start_sec"": 602.48, ""end_sec"": 606.399, ""text"": ""risks"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:04.399"", ""start_sec"": 604.399, ""end_sec"": 609.04, ""text"": ""now you must be curious to understand"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:06.399"", ""start_sec"": 606.399, ""end_sec"": 611.279, ""text"": ""the features of machine learning machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:09.040"", ""start_sec"": 609.04, ""end_sec"": 614.0, ""text"": ""learning uses the data to detect"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:11.279"", ""start_sec"": 611.279, ""end_sec"": 616.48, ""text"": ""patterns in a data set and adjust"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:14.000"", ""start_sec"": 614.0, ""end_sec"": 618.72, ""text"": ""program actions accordingly"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:16.480"", ""start_sec"": 616.48, ""end_sec"": 620.64, ""text"": ""pattern detection can be defined as the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:18.720"", ""start_sec"": 618.72, ""end_sec"": 623.2, ""text"": ""classification of data based on"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:20.640"", ""start_sec"": 620.64, ""end_sec"": 625.36, ""text"": ""knowledge already gained or on"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:23.200"", ""start_sec"": 623.2, ""end_sec"": 626.8, ""text"": ""statistical information extracted from"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:25.360"", ""start_sec"": 625.36, ""end_sec"": 628.64, ""text"": ""the patterns"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:26.800"", ""start_sec"": 626.8, ""end_sec"": 630.48, ""text"": ""it focuses on the development of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:28.640"", ""start_sec"": 628.64, ""end_sec"": 632.48, ""text"": ""computer programs that can teach"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:30.480"", ""start_sec"": 630.48, ""end_sec"": 634.56, ""text"": ""themselves to grow and change"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:32.480"", ""start_sec"": 632.48, ""end_sec"": 637.279, ""text"": ""when exposed to new data by using a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:34.560"", ""start_sec"": 634.56, ""end_sec"": 639.76, ""text"": ""method called reinforcement learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:37.279"", ""start_sec"": 637.279, ""end_sec"": 642.399, ""text"": ""it uses external feedback to teach the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:39.760"", ""start_sec"": 639.76, ""end_sec"": 644.88, ""text"": ""system to change its internal workings"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:42.399"", ""start_sec"": 642.399, ""end_sec"": 646.88, ""text"": ""in order to guess better next time"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:44.880"", ""start_sec"": 644.88, ""end_sec"": 649.6, ""text"": ""it enables computers to find hidden"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:46.880"", ""start_sec"": 646.88, ""end_sec"": 652.64, ""text"": ""insights using iterative algorithms"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:49.600"", ""start_sec"": 649.6, ""end_sec"": 655.12, ""text"": ""without being explicitly programmed"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:52.640"", ""start_sec"": 652.64, ""end_sec"": 657.519, ""text"": ""machine learning uses algorithms that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:55.120"", ""start_sec"": 655.12, ""end_sec"": 660.399, ""text"": ""learn from previous data to help produce"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:10:57.519"", ""start_sec"": 657.519, ""end_sec"": 662.64, ""text"": ""reliable and repeatable decisions it"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:00.399"", ""start_sec"": 660.399, ""end_sec"": 664.56, ""text"": ""automates analytical model building"", ""video_id"": ""SSE4M0gcmvE""}]",563.8389999999999,664.56
using the statistical and machine learning algorithms that tease patterns and relationships from data and express them as mathematicalequationslet's understand the different machine learningapproachesso what is the actual difference between traditional programming and machinelearningin traditionalprogrammingdata and is provided to thecomputerit processes them and gives theoutputhoweverthe machine learning approach is verydifferentin machinelearningalgorithms are applied on the given data andoutputthe result of the applied algorithm and calculations is a learning model that helps machine to learn from thedatain traditional programming you code the behavior of theprogrambut in machine learning you leave a lot of that to the machine to learn fromdatanow let's first understand the traditional programmingapproachtraditionally you would hard code the decision rules for a problem athandevaluate the results of the program and if the results weresatisfactorythe program would be deployed inproductionif the results were not asexpectedone would review theerrorschange the program and evaluate itagainthis iterative process continues till one gets the expectedresultwhat is the machine learningapproachin the new machine learningapproachthe decision rules are not hardcodedthe problem is solved by training a model with the training data in order to derive or learn an algorithm thatbest,SSE4M0gcmvE,00:11:02.640,"[{""timestamp"": ""00:11:02.640"", ""start_sec"": 662.64, ""end_sec"": 667.36, ""text"": ""using the statistical and machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:04.560"", ""start_sec"": 664.56, ""end_sec"": 670.24, ""text"": ""learning algorithms that tease patterns"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:07.360"", ""start_sec"": 667.36, ""end_sec"": 673.2, ""text"": ""and relationships from data and express"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:10.240"", ""start_sec"": 670.24, ""end_sec"": 675.12, ""text"": ""them as mathematical equations"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:13.200"", ""start_sec"": 673.2, ""end_sec"": 678.16, ""text"": ""let's understand the different machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:15.120"", ""start_sec"": 675.12, ""end_sec"": 678.16, ""text"": ""learning approaches"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:18.880"", ""start_sec"": 678.88, ""end_sec"": 683.519, ""text"": ""so what is the actual difference between"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:21.519"", ""start_sec"": 681.519, ""end_sec"": 686.56, ""text"": ""traditional programming and machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:23.519"", ""start_sec"": 683.519, ""end_sec"": 687.36, ""text"": ""learning in traditional programming data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:26.560"", ""start_sec"": 686.56, ""end_sec"": 690.32, ""text"": ""and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:27.360"", ""start_sec"": 687.36, ""end_sec"": 692.72, ""text"": ""is provided to the computer it processes"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:30.320"", ""start_sec"": 690.32, ""end_sec"": 694.56, ""text"": ""them and gives the output however the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:32.720"", ""start_sec"": 692.72, ""end_sec"": 697.36, ""text"": ""machine learning approach is very"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:34.560"", ""start_sec"": 694.56, ""end_sec"": 700.9590000000001, ""text"": ""different in machine learning algorithms"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:37.360"", ""start_sec"": 697.36, ""end_sec"": 703.36, ""text"": ""are applied on the given data and output"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:40.959"", ""start_sec"": 700.9590000000001, ""end_sec"": 706.0, ""text"": ""the result of the applied algorithm and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:43.360"", ""start_sec"": 703.36, ""end_sec"": 709.36, ""text"": ""calculations is a learning model that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:46.000"", ""start_sec"": 706.0, ""end_sec"": 711.68, ""text"": ""helps machine to learn from the data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:49.360"", ""start_sec"": 709.36, ""end_sec"": 714.32, ""text"": ""in traditional programming you code the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:51.680"", ""start_sec"": 711.68, ""end_sec"": 716.56, ""text"": ""behavior of the program but in machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:54.320"", ""start_sec"": 714.32, ""end_sec"": 719.12, ""text"": ""learning you leave a lot of that to the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:56.560"", ""start_sec"": 716.56, ""end_sec"": 720.56, ""text"": ""machine to learn from data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:11:59.120"", ""start_sec"": 719.12, ""end_sec"": 723.04, ""text"": ""now let's first understand the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:00.560"", ""start_sec"": 720.56, ""end_sec"": 725.04, ""text"": ""traditional programming approach"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:03.040"", ""start_sec"": 723.04, ""end_sec"": 727.68, ""text"": ""traditionally you would hard code the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:05.040"", ""start_sec"": 725.04, ""end_sec"": 730.32, ""text"": ""decision rules for a problem at hand"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:07.680"", ""start_sec"": 727.68, ""end_sec"": 732.24, ""text"": ""evaluate the results of the program and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:10.320"", ""start_sec"": 730.32, ""end_sec"": 735.279, ""text"": ""if the results were satisfactory the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:12.240"", ""start_sec"": 732.24, ""end_sec"": 737.68, ""text"": ""program would be deployed in production"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:15.279"", ""start_sec"": 735.279, ""end_sec"": 740.079, ""text"": ""if the results were not as expected one"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:17.680"", ""start_sec"": 737.68, ""end_sec"": 742.72, ""text"": ""would review the errors change the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:20.079"", ""start_sec"": 740.079, ""end_sec"": 745.279, ""text"": ""program and evaluate it again"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:22.720"", ""start_sec"": 742.72, ""end_sec"": 748.8, ""text"": ""this iterative process continues till"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:25.279"", ""start_sec"": 745.279, ""end_sec"": 751.2, ""text"": ""one gets the expected result"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:28.800"", ""start_sec"": 748.8, ""end_sec"": 753.12, ""text"": ""what is the machine learning approach in"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:31.200"", ""start_sec"": 751.2, ""end_sec"": 755.92, ""text"": ""the new machine learning approach the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:33.120"", ""start_sec"": 753.12, ""end_sec"": 758.24, ""text"": ""decision rules are not hard coded the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:35.920"", ""start_sec"": 755.92, ""end_sec"": 760.16, ""text"": ""problem is solved by training a model"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:38.240"", ""start_sec"": 758.24, ""end_sec"": 763.279, ""text"": ""with the training data in order to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:40.160"", ""start_sec"": 760.16, ""end_sec"": 765.76, ""text"": ""derive or learn an algorithm that best"", ""video_id"": ""SSE4M0gcmvE""}]",662.64,765.76
represents the relationship between the input and theoutputthis trained model is then evaluated against testdataif the results weresatisfactorythe model would be deployed inproductionand if the results are notsatisfactorythe training is repeated with somechangesmachine learningtechniquesmachine learning uses a number of theories and techniques from datasciencehere are some machine learningtechniquesclassification categorization clusteringtrendanalysisanomalydetectionvisualization and decisionmakinglet's look at thesetechniquesclassification is a technique in which the computer program learns from the data input given to it and then uses this learning to classify newobservationsclassification is used for predicting discreteresponsesclassification is used when we are training a model to predict qualitativetargetscategorization is a technique of organizing data into categories for its most effective and efficientuseit makes free text searches faster and provides a better userexperienceclustering is a technique of grouping a set of objects in such a way that objects in the same group are most similar to each other than to those in othergroupsit is basically a collection of objects on the basis of similarity and dissimilarity betweenthemtrend analysis is a technique aimed at projecting both current andfuture,SSE4M0gcmvE,00:12:43.279,"[{""timestamp"": ""00:12:43.279"", ""start_sec"": 763.279, ""end_sec"": 768.639, ""text"": ""represents the relationship between the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:45.760"", ""start_sec"": 765.76, ""end_sec"": 771.68, ""text"": ""input and the output this trained model"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:48.639"", ""start_sec"": 768.639, ""end_sec"": 773.8389999999999, ""text"": ""is then evaluated against test data if"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:51.680"", ""start_sec"": 771.68, ""end_sec"": 776.16, ""text"": ""the results were satisfactory the model"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:53.839"", ""start_sec"": 773.8389999999999, ""end_sec"": 778.16, ""text"": ""would be deployed in production and if"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:56.160"", ""start_sec"": 776.16, ""end_sec"": 781.92, ""text"": ""the results are not satisfactory the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:12:58.160"", ""start_sec"": 778.16, ""end_sec"": 785.36, ""text"": ""training is repeated with some changes"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:01.920"", ""start_sec"": 781.92, ""end_sec"": 785.36, ""text"": ""machine learning techniques"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:05.600"", ""start_sec"": 785.6, ""end_sec"": 789.36, ""text"": ""machine learning uses a number of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:07.600"", ""start_sec"": 787.6, ""end_sec"": 792.079, ""text"": ""theories and techniques from data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:09.360"", ""start_sec"": 789.36, ""end_sec"": 794.839, ""text"": ""science here are some machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:12.079"", ""start_sec"": 792.079, ""end_sec"": 798.48, ""text"": ""techniques classification"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:14.839"", ""start_sec"": 794.839, ""end_sec"": 801.44, ""text"": ""categorization clustering trend analysis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:18.480"", ""start_sec"": 798.48, ""end_sec"": 802.959, ""text"": ""anomaly detection visualization and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:21.440"", ""start_sec"": 801.44, ""end_sec"": 806.079, ""text"": ""decision making"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:22.959"", ""start_sec"": 802.959, ""end_sec"": 808.16, ""text"": ""let's look at these techniques"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:26.079"", ""start_sec"": 806.079, ""end_sec"": 810.16, ""text"": ""classification is a technique in which"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:28.160"", ""start_sec"": 808.16, ""end_sec"": 813.36, ""text"": ""the computer program learns from the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:30.160"", ""start_sec"": 810.16, ""end_sec"": 815.04, ""text"": ""data input given to it and then uses"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:33.360"", ""start_sec"": 813.36, ""end_sec"": 816.639, ""text"": ""this learning to classify new"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:35.040"", ""start_sec"": 815.04, ""end_sec"": 818.9590000000001, ""text"": ""observations"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:36.639"", ""start_sec"": 816.639, ""end_sec"": 821.8389999999999, ""text"": ""classification is used for predicting"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:38.959"", ""start_sec"": 818.9590000000001, ""end_sec"": 824.0, ""text"": ""discrete responses classification is"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:41.839"", ""start_sec"": 821.8389999999999, ""end_sec"": 828.36, ""text"": ""used when we are training a model to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:44.000"", ""start_sec"": 824.0, ""end_sec"": 830.399, ""text"": ""predict qualitative targets"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:48.360"", ""start_sec"": 828.36, ""end_sec"": 832.9590000000001, ""text"": ""categorization is a technique of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:50.399"", ""start_sec"": 830.399, ""end_sec"": 835.44, ""text"": ""organizing data into categories for its"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:52.959"", ""start_sec"": 832.9590000000001, ""end_sec"": 837.8389999999999, ""text"": ""most effective and efficient use"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:55.440"", ""start_sec"": 835.44, ""end_sec"": 840.959, ""text"": ""it makes free text searches faster and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:13:57.839"", ""start_sec"": 837.8389999999999, ""end_sec"": 843.279, ""text"": ""provides a better user experience"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:00.959"", ""start_sec"": 840.959, ""end_sec"": 845.12, ""text"": ""clustering is a technique of grouping a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:03.279"", ""start_sec"": 843.279, ""end_sec"": 847.199, ""text"": ""set of objects in such a way that"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:05.120"", ""start_sec"": 845.12, ""end_sec"": 849.6, ""text"": ""objects in the same group are most"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:07.199"", ""start_sec"": 847.199, ""end_sec"": 850.88, ""text"": ""similar to each other than to those in"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:09.600"", ""start_sec"": 849.6, ""end_sec"": 853.12, ""text"": ""other groups"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:10.880"", ""start_sec"": 850.88, ""end_sec"": 854.959, ""text"": ""it is basically a collection of objects"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:13.120"", ""start_sec"": 853.12, ""end_sec"": 858.0, ""text"": ""on the basis of similarity and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:14.959"", ""start_sec"": 854.959, ""end_sec"": 860.32, ""text"": ""dissimilarity between them"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:18.000"", ""start_sec"": 858.0, ""end_sec"": 862.24, ""text"": ""trend analysis is a technique aimed at"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:20.320"", ""start_sec"": 860.32, ""end_sec"": 864.639, ""text"": ""projecting both current and future"", ""video_id"": ""SSE4M0gcmvE""}]",763.279,864.639
movement of events through the use of time series dataanalysisit represents variations of low frequency in a timeseriesthe high and medium frequency fluctuations beingoutanomaly detection is a technique to identify cases that are unusual within data that is seeminglyhomogenousanomaly detection can be a key for solving intrusions by indicating a presence of intended or unintended inducedattacks defectsfaults and soonvisualization is a technique to present data in a pictorial or graphicalformatit enables decision makers to see analytics presentedvisuallywhen data is shown in the form ofpicturesit becomes easy for users to understanditdecision making is a technique or skill that provides you with the ability to influence managerialdecisionswith data as evidence for thosepossibilitiesnow i am sure you have a better understanding of the overview of machinelearningso let's look at some real-time applications of machinelearningartificial intelligence and machine learning are being increasingly used in variousfunctionssuch as imageprocessing roboticsdataminingvideogamestext analysis andhealthcarelet's look at each of them in moredetailsso what is imageprocessingit is a technique to convert an image into a digital format and perform some operations on it so as to induce an enhanced image or to extract some helpful information fromit,SSE4M0gcmvE,00:14:22.240,"[{""timestamp"": ""00:14:22.240"", ""start_sec"": 862.24, ""end_sec"": 867.279, ""text"": ""movement of events through the use of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:24.639"", ""start_sec"": 864.639, ""end_sec"": 869.44, ""text"": ""time series data analysis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:27.279"", ""start_sec"": 867.279, ""end_sec"": 872.399, ""text"": ""it represents variations of low"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:29.440"", ""start_sec"": 869.44, ""end_sec"": 876.399, ""text"": ""frequency in a time series the high and"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:32.399"", ""start_sec"": 872.399, ""end_sec"": 878.88, ""text"": ""medium frequency fluctuations being out"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:36.399"", ""start_sec"": 876.399, ""end_sec"": 881.519, ""text"": ""anomaly detection is a technique to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:38.880"", ""start_sec"": 878.88, ""end_sec"": 884.56, ""text"": ""identify cases that are unusual within"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:41.519"", ""start_sec"": 881.519, ""end_sec"": 886.639, ""text"": ""data that is seemingly homogenous"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:44.560"", ""start_sec"": 884.56, ""end_sec"": 888.72, ""text"": ""anomaly detection can be a key for"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:46.639"", ""start_sec"": 886.639, ""end_sec"": 891.36, ""text"": ""solving intrusions by indicating a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:48.720"", ""start_sec"": 888.72, ""end_sec"": 895.639, ""text"": ""presence of intended or unintended"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:51.360"", ""start_sec"": 891.36, ""end_sec"": 898.0, ""text"": ""induced attacks defects faults and so on"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:55.639"", ""start_sec"": 895.639, ""end_sec"": 901.68, ""text"": ""visualization is a technique to present"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:14:58.000"", ""start_sec"": 898.0, ""end_sec"": 903.76, ""text"": ""data in a pictorial or graphical format"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:01.680"", ""start_sec"": 901.68, ""end_sec"": 906.32, ""text"": ""it enables decision makers to see"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:03.760"", ""start_sec"": 903.76, ""end_sec"": 908.399, ""text"": ""analytics presented visually"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:06.320"", ""start_sec"": 906.32, ""end_sec"": 910.88, ""text"": ""when data is shown in the form of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:08.399"", ""start_sec"": 908.399, ""end_sec"": 912.48, ""text"": ""pictures it becomes easy for users to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:10.880"", ""start_sec"": 910.88, ""end_sec"": 914.639, ""text"": ""understand it"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:12.480"", ""start_sec"": 912.48, ""end_sec"": 916.56, ""text"": ""decision making is a technique or skill"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:14.639"", ""start_sec"": 914.639, ""end_sec"": 919.6, ""text"": ""that provides you with the ability to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:16.560"", ""start_sec"": 916.56, ""end_sec"": 922.8, ""text"": ""influence managerial decisions with data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:19.600"", ""start_sec"": 919.6, ""end_sec"": 924.72, ""text"": ""as evidence for those possibilities"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:22.800"", ""start_sec"": 922.8, ""end_sec"": 926.8, ""text"": ""now i am sure you have a better"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:24.720"", ""start_sec"": 924.72, ""end_sec"": 929.12, ""text"": ""understanding of the overview of machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:26.800"", ""start_sec"": 926.8, ""end_sec"": 931.68, ""text"": ""learning so let's look at some real-time"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:29.120"", ""start_sec"": 929.12, ""end_sec"": 933.8389999999999, ""text"": ""applications of machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:31.680"", ""start_sec"": 931.68, ""end_sec"": 936.16, ""text"": ""artificial intelligence and machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:33.839"", ""start_sec"": 933.8389999999999, ""end_sec"": 938.32, ""text"": ""learning are being increasingly used in"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:36.160"", ""start_sec"": 936.16, ""end_sec"": 940.399, ""text"": ""various functions such as image"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:38.320"", ""start_sec"": 938.32, ""end_sec"": 944.079, ""text"": ""processing robotics"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:40.399"", ""start_sec"": 940.399, ""end_sec"": 946.399, ""text"": ""data mining video games text analysis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:44.079"", ""start_sec"": 944.079, ""end_sec"": 948.32, ""text"": ""and healthcare let's look at each of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:46.399"", ""start_sec"": 946.399, ""end_sec"": 951.199, ""text"": ""them in more details"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:48.320"", ""start_sec"": 948.32, ""end_sec"": 953.279, ""text"": ""so what is image processing it is a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:51.199"", ""start_sec"": 951.199, ""end_sec"": 955.519, ""text"": ""technique to convert an image into a"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:53.279"", ""start_sec"": 953.279, ""end_sec"": 958.16, ""text"": ""digital format and perform some"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:55.519"", ""start_sec"": 955.519, ""end_sec"": 960.56, ""text"": ""operations on it so as to induce an"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:15:58.160"", ""start_sec"": 958.16, ""end_sec"": 962.8, ""text"": ""enhanced image or to extract some"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:00.560"", ""start_sec"": 960.56, ""end_sec"": 964.72, ""text"": ""helpful information from it"", ""video_id"": ""SSE4M0gcmvE""}]",862.24,964.72
let's look at some of the examples of imageprocessingfacebook does automatic face tagging by recognizing a face from a previous user's taggedphotosanother example is optional characterrecognitionwhich scans printed docs to digitize thetextself-driving cars are another big example of imageprocessingautopilot is an optional drive system for teslacarswhen autopilot isengagedcars canself-steeradjustspeeddetect nearbyobstaclesapply the brakes andparknow let's see how robotics uses machinelearningrobots are machines that can be used to do certainjobssome of the examples of robotics are where a humanoid robot can read the emotions of humanbeingsor an industrial robot is used for assembling and manufacturingproductsso let's look at some real-time applications of machinelearninglet's see what data miningisit is the method of analyzing hidden patterns indatalet's look at some of the applications of dataminingit is used for anomalydetectionto detect credit card fraud and to determine which transactions vary from usual purchasingpatternsit is also used in market basketanalysiswhich is used to detect which items are often boughttogetherit can be used forgroupingwhere it classifies users based on theirprofilesmachine learning is also applied in many,SSE4M0gcmvE,00:16:02.800,"[{""timestamp"": ""00:16:02.800"", ""start_sec"": 962.8, ""end_sec"": 966.399, ""text"": ""let's look at some of the examples of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:04.720"", ""start_sec"": 964.72, ""end_sec"": 968.88, ""text"": ""image processing"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:06.399"", ""start_sec"": 966.399, ""end_sec"": 970.959, ""text"": ""facebook does automatic face tagging by"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:08.880"", ""start_sec"": 968.88, ""end_sec"": 974.0, ""text"": ""recognizing a face from a previous"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:10.959"", ""start_sec"": 970.959, ""end_sec"": 975.839, ""text"": ""user's tagged photos another example is"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:14.000"", ""start_sec"": 974.0, ""end_sec"": 979.279, ""text"": ""optional character recognition which"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:15.839"", ""start_sec"": 975.839, ""end_sec"": 981.04, ""text"": ""scans printed docs to digitize the text"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:19.279"", ""start_sec"": 979.279, ""end_sec"": 983.36, ""text"": ""self-driving cars are another big"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:21.040"", ""start_sec"": 981.04, ""end_sec"": 986.079, ""text"": ""example of image processing"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:23.360"", ""start_sec"": 983.36, ""end_sec"": 987.8389999999999, ""text"": ""autopilot is an optional drive system"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:26.079"", ""start_sec"": 986.079, ""end_sec"": 990.399, ""text"": ""for tesla cars"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:27.839"", ""start_sec"": 987.8389999999999, ""end_sec"": 994.16, ""text"": ""when autopilot is engaged cars can"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:30.399"", ""start_sec"": 990.399, ""end_sec"": 997.68, ""text"": ""self-steer adjust speed detect nearby"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:34.160"", ""start_sec"": 994.16, ""end_sec"": 1000.079, ""text"": ""obstacles apply the brakes and park"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:37.680"", ""start_sec"": 997.68, ""end_sec"": 1001.04, ""text"": ""now let's see how robotics uses machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:40.079"", ""start_sec"": 1000.079, ""end_sec"": 1003.12, ""text"": ""learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:41.040"", ""start_sec"": 1001.04, ""end_sec"": 1004.72, ""text"": ""robots are machines that can be used to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:43.120"", ""start_sec"": 1003.12, ""end_sec"": 1007.199, ""text"": ""do certain jobs"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:44.720"", ""start_sec"": 1004.72, ""end_sec"": 1009.199, ""text"": ""some of the examples of robotics are"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:47.199"", ""start_sec"": 1007.199, ""end_sec"": 1012.079, ""text"": ""where a humanoid robot can read the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:49.199"", ""start_sec"": 1009.199, ""end_sec"": 1014.079, ""text"": ""emotions of human beings or"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:52.079"", ""start_sec"": 1012.079, ""end_sec"": 1017.04, ""text"": ""an industrial robot is used for"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:54.079"", ""start_sec"": 1014.079, ""end_sec"": 1018.88, ""text"": ""assembling and manufacturing products"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:57.040"", ""start_sec"": 1017.04, ""end_sec"": 1021.36, ""text"": ""so let's look at some real-time"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:16:58.880"", ""start_sec"": 1018.88, ""end_sec"": 1024.88, ""text"": ""applications of machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:01.360"", ""start_sec"": 1021.36, ""end_sec"": 1027.199, ""text"": ""let's see what data mining is it is the"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:04.880"", ""start_sec"": 1024.88, ""end_sec"": 1028.16, ""text"": ""method of analyzing hidden patterns in"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:07.199"", ""start_sec"": 1027.199, ""end_sec"": 1030.16, ""text"": ""data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:08.160"", ""start_sec"": 1028.16, ""end_sec"": 1031.679, ""text"": ""let's look at some of the applications"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:10.160"", ""start_sec"": 1030.16, ""end_sec"": 1033.919, ""text"": ""of data mining"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:11.679"", ""start_sec"": 1031.679, ""end_sec"": 1036.0, ""text"": ""it is used for anomaly detection to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:13.919"", ""start_sec"": 1033.919, ""end_sec"": 1038.72, ""text"": ""detect credit card fraud and to"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:16.000"", ""start_sec"": 1036.0, ""end_sec"": 1041.839, ""text"": ""determine which transactions vary from"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:18.720"", ""start_sec"": 1038.72, ""end_sec"": 1044.16, ""text"": ""usual purchasing patterns"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:21.839"", ""start_sec"": 1041.839, ""end_sec"": 1046.799, ""text"": ""it is also used in market basket"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:24.160"", ""start_sec"": 1044.16, ""end_sec"": 1050.559, ""text"": ""analysis which is used to detect which"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:26.799"", ""start_sec"": 1046.799, ""end_sec"": 1050.559, ""text"": ""items are often bought together"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:31.679"", ""start_sec"": 1051.679, ""end_sec"": 1058.72, ""text"": ""it can be used for grouping where it"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:33.679"", ""start_sec"": 1053.679, ""end_sec"": 1058.72, ""text"": ""classifies users based on their profiles"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:38.799"", ""start_sec"": 1058.799, ""end_sec"": 1063.28, ""text"": ""machine learning is also applied in many"", ""video_id"": ""SSE4M0gcmvE""}]",962.8,1063.28
video games in order to give predictions based ondatain a pokemon gobattlethere is a lot of data to take into account to correctly predict the winner of abattleand this is where machine learning becomesusefula machine learning classifier will predict the result of the match based on thisdatalet's move on to one of the most popular applications of machinelearningwhich is textanalysisit is the automated process of obtaining information fromtextone example of text analysis is spamfilteringwhich is used to detect spam inemailsanother example is sentimentalanalysiswhich is used for classifying an opinion aspositivenegative orneutralit detects public sentiment in twitter feed or filters customercomplaintsit is also used for informationextractionsuch as extracting specificdata addresskeyword orentitiesthere are many applications of machine learning in the healthcareindustryidentifying disease anddiagnosisdrug discovery andmanufacturingmedical imagingdiagnosisand soonsome of the companies that use machine learning have revolutionized the health care industry aregoogledeepmind healthbiobeatshealth fidelity and ginger dot ioyou,SSE4M0gcmvE,00:17:41.039,"[{""timestamp"": ""00:17:41.039"", ""start_sec"": 1061.039, ""end_sec"": 1066.799, ""text"": ""video games in order to give predictions"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:43.280"", ""start_sec"": 1063.28, ""end_sec"": 1068.64, ""text"": ""based on data in a pokemon go battle"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:46.799"", ""start_sec"": 1066.799, ""end_sec"": 1070.48, ""text"": ""there is a lot of data to take into"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:48.640"", ""start_sec"": 1068.64, ""end_sec"": 1071.76, ""text"": ""account to correctly predict the winner"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:50.480"", ""start_sec"": 1070.48, ""end_sec"": 1073.6, ""text"": ""of a battle"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:51.760"", ""start_sec"": 1071.76, ""end_sec"": 1076.0, ""text"": ""and this is where machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:53.600"", ""start_sec"": 1073.6, ""end_sec"": 1078.4, ""text"": ""becomes useful a machine learning"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:56.000"", ""start_sec"": 1076.0, ""end_sec"": 1081.28, ""text"": ""classifier will predict the result of"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:17:58.400"", ""start_sec"": 1078.4, ""end_sec"": 1083.52, ""text"": ""the match based on this data"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:01.280"", ""start_sec"": 1081.28, ""end_sec"": 1085.44, ""text"": ""let's move on to one of the most popular"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:03.520"", ""start_sec"": 1083.52, ""end_sec"": 1087.36, ""text"": ""applications of machine learning which"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:05.440"", ""start_sec"": 1085.44, ""end_sec"": 1089.919, ""text"": ""is text analysis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:07.360"", ""start_sec"": 1087.36, ""end_sec"": 1091.84, ""text"": ""it is the automated process of obtaining"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:09.919"", ""start_sec"": 1089.919, ""end_sec"": 1094.64, ""text"": ""information from text"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:11.840"", ""start_sec"": 1091.84, ""end_sec"": 1097.6, ""text"": ""one example of text analysis is spam"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:14.640"", ""start_sec"": 1094.64, ""end_sec"": 1099.039, ""text"": ""filtering which is used to detect spam"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:17.600"", ""start_sec"": 1097.6, ""end_sec"": 1101.44, ""text"": ""in emails"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:19.039"", ""start_sec"": 1099.039, ""end_sec"": 1104.16, ""text"": ""another example is sentimental analysis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:21.440"", ""start_sec"": 1101.44, ""end_sec"": 1106.4, ""text"": ""which is used for classifying an opinion"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:24.160"", ""start_sec"": 1104.16, ""end_sec"": 1108.799, ""text"": ""as positive negative or neutral it"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:26.400"", ""start_sec"": 1106.4, ""end_sec"": 1111.36, ""text"": ""detects public sentiment in twitter feed"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:28.799"", ""start_sec"": 1108.799, ""end_sec"": 1113.28, ""text"": ""or filters customer complaints"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:31.360"", ""start_sec"": 1111.36, ""end_sec"": 1116.32, ""text"": ""it is also used for information"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:33.280"", ""start_sec"": 1113.28, ""end_sec"": 1120.88, ""text"": ""extraction such as extracting specific"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:36.320"", ""start_sec"": 1116.32, ""end_sec"": 1120.88, ""text"": ""data address keyword or entities"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:41.200"", ""start_sec"": 1121.2, ""end_sec"": 1125.52, ""text"": ""there are many applications of machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:43.280"", ""start_sec"": 1123.28, ""end_sec"": 1128.72, ""text"": ""learning in the healthcare industry"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:45.520"", ""start_sec"": 1125.52, ""end_sec"": 1131.6, ""text"": ""identifying disease and diagnosis"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:48.720"", ""start_sec"": 1128.72, ""end_sec"": 1134.4, ""text"": ""drug discovery and manufacturing medical"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:51.600"", ""start_sec"": 1131.6, ""end_sec"": 1136.32, ""text"": ""imaging diagnosis and so on"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:54.400"", ""start_sec"": 1134.4, ""end_sec"": 1138.48, ""text"": ""some of the companies that use machine"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:56.320"", ""start_sec"": 1136.32, ""end_sec"": 1141.2, ""text"": ""learning have revolutionized the health"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:18:58.480"", ""start_sec"": 1138.48, ""end_sec"": 1142.16, ""text"": ""care industry are google deep mind"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:19:01.200"", ""start_sec"": 1141.2, ""end_sec"": 1146.32, ""text"": ""health"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:19:02.160"", ""start_sec"": 1142.16, ""end_sec"": 1148.43, ""text"": ""bio beats health fidelity and ginger dot"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:19:06.320"", ""start_sec"": 1146.32, ""end_sec"": 1152.88, ""text"": ""io"", ""video_id"": ""SSE4M0gcmvE""}, {""timestamp"": ""00:19:12.880"", ""start_sec"": 1152.88, ""end_sec"": 1154.96, ""text"": ""you"", ""video_id"": ""SSE4M0gcmvE""}]",1061.039,1154.96
hello and welcome to the session on deeplearningmy name is mohan and in this video we are going to talk about what deep learning is allaboutsome of you may be already familiar with the imagerecognitionhow does image recognitionworkyou can train this application or your machine to recognize whether a given image is a cat or adogand this is how it works at a very highlevelit uses artificial neuralnetworkit is trained with some known images and during the training it is told if it is recognizing correctly ornotand then when new images aresubmittedit recognizescorrectlybased on theaccuracyofcourseso a little quick understanding about artificial neuralnetworksso this is the way itdoesis you provide a lot of trainingdataalso known as labeleddataforexamplein this case these are the images ofdogsand the network extracts some features that makes a dog a dogrightso that is known as feature extraction and based onthatwhen you submit a new image ofdogthe basic features remain pretty much thesameit may be a completely differentimagebut the features of a dog still remain pretty much the same in various differentimageslet's say compared to acatand that's the way artificial neural networkworkswe'll go into details ofthis uhveryshortlyand once the training is done with trainingdatawe then test it with some test datatoowhich is basically completely new data which the system has not seenbeforeunlike the trainingdataand then we find out whether it ispredicting,FbxTVRfQFuI,00:00:03.040,"[{""timestamp"": ""00:00:03.040"", ""start_sec"": 3.04, ""end_sec"": 8.96, ""text"": ""hello and welcome to the session on deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:05.920"", ""start_sec"": 5.92, ""end_sec"": 11.599, ""text"": ""learning my name is mohan and in this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:08.960"", ""start_sec"": 8.96, ""end_sec"": 14.08, ""text"": ""video we are going to talk about what"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:11.599"", ""start_sec"": 11.599, ""end_sec"": 16.16, ""text"": ""deep learning is all about some of you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:14.080"", ""start_sec"": 14.08, ""end_sec"": 19.52, ""text"": ""may be already familiar with the image"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:16.160"", ""start_sec"": 16.16, ""end_sec"": 22.64, ""text"": ""recognition how does image recognition"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:19.520"", ""start_sec"": 19.52, ""end_sec"": 25.68, ""text"": ""work you can train this application or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:22.640"", ""start_sec"": 22.64, ""end_sec"": 28.08, ""text"": ""your machine to recognize whether a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:25.680"", ""start_sec"": 25.68, ""end_sec"": 30.08, ""text"": ""given image is a cat or a dog and this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:28.080"", ""start_sec"": 28.08, ""end_sec"": 32.239, ""text"": ""is how it works at a very high level it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:30.080"", ""start_sec"": 30.08, ""end_sec"": 34.48, ""text"": ""uses artificial neural network it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:32.239"", ""start_sec"": 32.239, ""end_sec"": 36.559, ""text"": ""trained with some known images and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:34.480"", ""start_sec"": 34.48, ""end_sec"": 38.96, ""text"": ""during the training it is told if it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:36.559"", ""start_sec"": 36.559, ""end_sec"": 41.04, ""text"": ""recognizing correctly or not and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:38.960"", ""start_sec"": 38.96, ""end_sec"": 42.96, ""text"": ""when new images are submitted it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:41.040"", ""start_sec"": 41.04, ""end_sec"": 45.68, ""text"": ""recognizes correctly based on the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:42.960"", ""start_sec"": 42.96, ""end_sec"": 47.6, ""text"": ""accuracy of course so a little quick"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:45.680"", ""start_sec"": 45.68, ""end_sec"": 50.48, ""text"": ""understanding about artificial neural"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:47.600"", ""start_sec"": 47.6, ""end_sec"": 53.6, ""text"": ""networks so this is the way it does is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:50.480"", ""start_sec"": 50.48, ""end_sec"": 56.0, ""text"": ""you provide a lot of training data also"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:53.600"", ""start_sec"": 53.6, ""end_sec"": 59.359, ""text"": ""known as labeled data for example in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:56.000"", ""start_sec"": 56.0, ""end_sec"": 62.64, ""text"": ""this case these are the images of dogs"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:00:59.359"", ""start_sec"": 59.359, ""end_sec"": 65.6, ""text"": ""and the network extracts some features"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:02.640"", ""start_sec"": 62.64, ""end_sec"": 68.64, ""text"": ""that makes a dog a dog right so that is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:05.600"", ""start_sec"": 65.6, ""end_sec"": 71.76, ""text"": ""known as feature extraction and based on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:08.640"", ""start_sec"": 68.64, ""end_sec"": 73.76, ""text"": ""that when you submit a new image of dog"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:11.760"", ""start_sec"": 71.76, ""end_sec"": 75.119, ""text"": ""the basic features remain pretty much"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:13.760"", ""start_sec"": 73.76, ""end_sec"": 77.759, ""text"": ""the same it may be a completely"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:15.119"", ""start_sec"": 75.119, ""end_sec"": 81.28, ""text"": ""different image but the features of a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:17.759"", ""start_sec"": 77.759, ""end_sec"": 83.2, ""text"": ""dog still remain pretty much the same in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:21.280"", ""start_sec"": 81.28, ""end_sec"": 85.68, ""text"": ""various different images let's say"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:23.200"", ""start_sec"": 83.2, ""end_sec"": 88.0, ""text"": ""compared to a cat and that's the way"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:25.680"", ""start_sec"": 85.68, ""end_sec"": 90.479, ""text"": ""artificial neural network works we'll go"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:28.000"", ""start_sec"": 88.0, ""end_sec"": 92.24000000000001, ""text"": ""into details of this uh very shortly and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:30.479"", ""start_sec"": 90.479, ""end_sec"": 95.119, ""text"": ""once the training is done with training"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:32.240"", ""start_sec"": 92.24000000000001, ""end_sec"": 97.439, ""text"": ""data we then test it with some test data"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:35.119"", ""start_sec"": 95.119, ""end_sec"": 99.84, ""text"": ""too which is basically completely new"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:37.439"", ""start_sec"": 97.439, ""end_sec"": 102.24000000000001, ""text"": ""data which the system has not seen"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:39.840"", ""start_sec"": 99.84, ""end_sec"": 103.92, ""text"": ""before unlike the training data and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:42.240"", ""start_sec"": 102.24000000000001, ""end_sec"": 106.56, ""text"": ""we find out whether it is predicting"", ""video_id"": ""FbxTVRfQFuI""}]",3.04,106.56
correctly ornotthereby we know whether the training is complete or it needs moretrainingso that's not a very high level artificial neural networkworksso this is what we are going to talk abouttodayour agenda looks something likethiswhat is deeplearningwhy do we need deeplearningand then what are the applications of deeplearningone of the maincomponentsthe secret sauce in deeplearningis neuralnetworksso we're going to talk about what is neural network and how it works and some of itscomponents likeforexamplethe activationfunctionthe gradient descent and so on and soforthsothatas a part of working of a neuralnetworkwe will go into little bit more details how this whole thingworkssowithout much furtheradolet's getstartedso deep learning is considered to be a part of machinelearningso this diagram very nicely depicts what deep learning is at a very highlevelyou have the all-encompassing artificialintelligencewhich is more a concept rather than a technology or a technicalconceptrightso itisit's more of a concept at a very highlevelartificial intelligence under the herd is actually machinelearningand deep learning and machine learning is a broaderconceptyou cansayor a broadertechnologyand deep learning is a subset of machinelearningthe primary difference between machine learning and deep learning is that deep learning uses neural networks and it is suitable for handling large amounts of unstructured data and the last but notleastone of,FbxTVRfQFuI,00:01:43.920,"[{""timestamp"": ""00:01:43.920"", ""start_sec"": 103.92, ""end_sec"": 109.28, ""text"": ""correctly or not thereby we know whether"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:46.560"", ""start_sec"": 106.56, ""end_sec"": 110.799, ""text"": ""the training is complete or it needs"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:49.280"", ""start_sec"": 109.28, ""end_sec"": 113.119, ""text"": ""more training so that's not a very high"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:50.799"", ""start_sec"": 110.799, ""end_sec"": 115.03999999999999, ""text"": ""level artificial neural network works so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:53.119"", ""start_sec"": 113.119, ""end_sec"": 117.119, ""text"": ""this is what we are going to talk about"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:55.040"", ""start_sec"": 115.03999999999999, ""end_sec"": 119.2, ""text"": ""today our agenda looks something like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:57.119"", ""start_sec"": 117.119, ""end_sec"": 120.799, ""text"": ""this what is deep learning why do we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:01:59.200"", ""start_sec"": 119.2, ""end_sec"": 123.04, ""text"": ""need deep learning and then what are the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:00.799"", ""start_sec"": 120.799, ""end_sec"": 125.92, ""text"": ""applications of deep learning one of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:03.040"", ""start_sec"": 123.04, ""end_sec"": 128.239, ""text"": ""main components the secret sauce in deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:05.920"", ""start_sec"": 125.92, ""end_sec"": 129.599, ""text"": ""learning is neural networks so we're"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:08.239"", ""start_sec"": 128.239, ""end_sec"": 130.879, ""text"": ""going to talk about what is neural"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:09.599"", ""start_sec"": 129.599, ""end_sec"": 132.879, ""text"": ""network and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:10.879"", ""start_sec"": 130.879, ""end_sec"": 135.52, ""text"": ""how it works and some of its components"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:12.879"", ""start_sec"": 132.879, ""end_sec"": 137.44, ""text"": ""like for example the activation function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:15.520"", ""start_sec"": 135.52, ""end_sec"": 140.16, ""text"": ""the gradient descent and so on and so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:17.440"", ""start_sec"": 137.44, ""end_sec"": 141.68, ""text"": ""forth so that as a part of working of a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:20.160"", ""start_sec"": 140.16, ""end_sec"": 143.52, ""text"": ""neural network we will go into little"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:21.680"", ""start_sec"": 141.68, ""end_sec"": 146.72, ""text"": ""bit more details how this whole thing"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:23.520"", ""start_sec"": 143.52, ""end_sec"": 149.36, ""text"": ""works so without much further ado let's"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:26.720"", ""start_sec"": 146.72, ""end_sec"": 151.52, ""text"": ""get started so deep learning is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:29.360"", ""start_sec"": 149.36, ""end_sec"": 154.07999999999998, ""text"": ""considered to be a part of machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:31.520"", ""start_sec"": 151.52, ""end_sec"": 156.72, ""text"": ""learning so this diagram very nicely"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:34.080"", ""start_sec"": 154.07999999999998, ""end_sec"": 159.599, ""text"": ""depicts what deep learning is at a very"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:36.720"", ""start_sec"": 156.72, ""end_sec"": 162.48, ""text"": ""high level you have the all-encompassing"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:39.599"", ""start_sec"": 159.599, ""end_sec"": 165.84, ""text"": ""artificial intelligence which is more a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:42.480"", ""start_sec"": 162.48, ""end_sec"": 167.68, ""text"": ""concept rather than a technology or a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:45.840"", ""start_sec"": 165.84, ""end_sec"": 169.28, ""text"": ""technical concept right so it is it's"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:47.680"", ""start_sec"": 167.68, ""end_sec"": 171.44, ""text"": ""more of a concept at a very high level"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:49.280"", ""start_sec"": 169.28, ""end_sec"": 173.2, ""text"": ""artificial intelligence under the herd"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:51.440"", ""start_sec"": 171.44, ""end_sec"": 175.36, ""text"": ""is actually machine learning and deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:53.200"", ""start_sec"": 173.2, ""end_sec"": 178.56, ""text"": ""learning and machine learning is a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:55.360"", ""start_sec"": 175.36, ""end_sec"": 181.84, ""text"": ""broader concept you can say or a broader"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:02:58.560"", ""start_sec"": 178.56, ""end_sec"": 183.28, ""text"": ""technology and deep learning is a subset"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:01.840"", ""start_sec"": 181.84, ""end_sec"": 185.44, ""text"": ""of machine learning the primary"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:03.280"", ""start_sec"": 183.28, ""end_sec"": 187.92, ""text"": ""difference between machine learning and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:05.440"", ""start_sec"": 185.44, ""end_sec"": 191.519, ""text"": ""deep learning is that deep learning uses"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:07.920"", ""start_sec"": 187.92, ""end_sec"": 194.08, ""text"": ""neural networks and it is suitable for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:11.519"", ""start_sec"": 191.519, ""end_sec"": 196.48, ""text"": ""handling large amounts of unstructured"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:14.080"", ""start_sec"": 194.08, ""end_sec"": 198.07999999999998, ""text"": ""data and the last but not least one of"", ""video_id"": ""FbxTVRfQFuI""}]",103.92,198.07999999999998
the major differences between machine learning and deep learning is that in machinelearningthe featureextractionor the featureengineeringis done by the data scientistsmanuallybut in deeplearningsince we use neuralnetworksthe feature engineering happensautomaticallyso that's a little bit of a quick difference between machine learning and deeplearningand this diagram very nicely depicts the relation between artificialintelligencemachine learning and deeplearningnow why do we need deeplearningmachine learning was there for quite some time and it can do a lot ofstuffthat probably what deep learning candobut it's not very good at handling large amounts of unstructured data likeimagesvoice or eventextfor thatmatterso traditional machine learning is not that very good at doingthistraditional machine learning can handle large amounts of structureddatabut when it comes to unstructured data it's a bigchallengeso that is one of the key differentiators for deeplearningso that is number one and increasingly for artificialintelligencewe need image recognition and we need toprocessanalyze images andvoicethat's the reason deep learning isrequiredcomparedtolet'ssaytraditional machinelearningit can also perform complexalgorithmsmore complexthanlet'ssaywhat machine learning candoand it can achieve best performance with the large amounts ofdataso the more you have thedatalet's say reference data or labeldatathe better thesystem,FbxTVRfQFuI,00:03:16.480,"[{""timestamp"": ""00:03:16.480"", ""start_sec"": 196.48, ""end_sec"": 199.599, ""text"": ""the major differences between machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:18.080"", ""start_sec"": 198.07999999999998, ""end_sec"": 202.07999999999998, ""text"": ""learning and deep learning is that in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:19.599"", ""start_sec"": 199.599, ""end_sec"": 204.64, ""text"": ""machine learning the feature extraction"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:22.080"", ""start_sec"": 202.07999999999998, ""end_sec"": 206.959, ""text"": ""or the feature engineering is done by"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:24.640"", ""start_sec"": 204.64, ""end_sec"": 209.28, ""text"": ""the data scientists manually but in deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:26.959"", ""start_sec"": 206.959, ""end_sec"": 210.799, ""text"": ""learning since we use neural networks"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:29.280"", ""start_sec"": 209.28, ""end_sec"": 212.72, ""text"": ""the feature engineering happens"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:30.799"", ""start_sec"": 210.799, ""end_sec"": 214.72, ""text"": ""automatically so that's a little bit of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:32.720"", ""start_sec"": 212.72, ""end_sec"": 216.0, ""text"": ""a quick difference between machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:34.720"", ""start_sec"": 214.72, ""end_sec"": 218.159, ""text"": ""learning and deep learning and this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:36.000"", ""start_sec"": 216.0, ""end_sec"": 220.0, ""text"": ""diagram very nicely depicts the relation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:38.159"", ""start_sec"": 218.159, ""end_sec"": 222.239, ""text"": ""between artificial intelligence machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:40.000"", ""start_sec"": 220.0, ""end_sec"": 224.31900000000002, ""text"": ""learning and deep learning now why do we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:42.239"", ""start_sec"": 222.239, ""end_sec"": 227.04, ""text"": ""need deep learning machine learning was"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:44.319"", ""start_sec"": 224.31900000000002, ""end_sec"": 229.12, ""text"": ""there for quite some time and it can do"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:47.040"", ""start_sec"": 227.04, ""end_sec"": 231.599, ""text"": ""a lot of stuff that probably what deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:49.120"", ""start_sec"": 229.12, ""end_sec"": 233.68, ""text"": ""learning can do but it's not very good"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:51.599"", ""start_sec"": 231.599, ""end_sec"": 237.2, ""text"": ""at handling large amounts of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:53.680"", ""start_sec"": 233.68, ""end_sec"": 239.92000000000002, ""text"": ""unstructured data like images voice or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:57.200"", ""start_sec"": 237.2, ""end_sec"": 241.92, ""text"": ""even text for that matter so traditional"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:03:59.920"", ""start_sec"": 239.92000000000002, ""end_sec"": 243.519, ""text"": ""machine learning is not that very good"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:01.920"", ""start_sec"": 241.92, ""end_sec"": 245.04, ""text"": ""at doing this traditional machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:03.519"", ""start_sec"": 243.519, ""end_sec"": 247.04, ""text"": ""learning can handle large amounts of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:05.040"", ""start_sec"": 245.04, ""end_sec"": 249.12, ""text"": ""structured data but when it comes to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:07.040"", ""start_sec"": 247.04, ""end_sec"": 250.48, ""text"": ""unstructured data it's a big challenge"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:09.120"", ""start_sec"": 249.12, ""end_sec"": 252.56, ""text"": ""so that is one of the key"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:10.480"", ""start_sec"": 250.48, ""end_sec"": 255.519, ""text"": ""differentiators for deep learning so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:12.560"", ""start_sec"": 252.56, ""end_sec"": 258.32, ""text"": ""that is number one and increasingly for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:15.519"", ""start_sec"": 255.519, ""end_sec"": 260.4, ""text"": ""artificial intelligence we need image"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:18.320"", ""start_sec"": 258.32, ""end_sec"": 262.32, ""text"": ""recognition and we need to process"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:20.400"", ""start_sec"": 260.4, ""end_sec"": 263.68, ""text"": ""analyze images and voice that's the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:22.320"", ""start_sec"": 262.32, ""end_sec"": 265.52, ""text"": ""reason deep learning is required"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:23.680"", ""start_sec"": 263.68, ""end_sec"": 267.84, ""text"": ""compared to let's say traditional"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:25.520"", ""start_sec"": 265.52, ""end_sec"": 271.199, ""text"": ""machine learning it can also perform"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:27.840"", ""start_sec"": 267.84, ""end_sec"": 273.12, ""text"": ""complex algorithms more complex than"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:31.199"", ""start_sec"": 271.199, ""end_sec"": 275.919, ""text"": ""let's say what machine learning can do"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:33.120"", ""start_sec"": 273.12, ""end_sec"": 278.0, ""text"": ""and it can achieve best performance with"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:35.919"", ""start_sec"": 275.919, ""end_sec"": 279.919, ""text"": ""the large amounts of data so the more"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:38.000"", ""start_sec"": 278.0, ""end_sec"": 282.8, ""text"": ""you have the data let's say reference"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:39.919"", ""start_sec"": 279.919, ""end_sec"": 284.639, ""text"": ""data or label data the better the system"", ""video_id"": ""FbxTVRfQFuI""}]",196.48,284.639
willdobecause the training process will be that muchbetterandlast but notleastwith deep learning you can really avoid the manual process of featureextractionthose are some of the reasons why we need deeplearningsome of the applications of deeplearningdeep learning has made major inroads and it is a major area in which deep learning is applied is healthcare and withinhealthcareparticularlyoncologywhich is basically cancer relatedstuffone of the issues with cancer is that a lot of cancers today arecurablethey can becuredthey are detected early on and the challenge with that is when a diagnostics isperformedlet's say an image has been taken of a patient to detect whether there is cancer ornotyou need a specialist to look at the image and determine whether itisthe patient is fine or there is any onset ofcancerand the number of specialists arelimitedso if we use deeplearningif we use automation here or if we use artificial intelligenceherethen the systemcanwith a certain amount of the good amount ofaccuracydetermine whether a particular patient is having cancer ornotso the prediction or the detection process of a disease like cancer can beexpeditedthe detection process can beexpeditedcan befasterwithout really waiting for aspecialistwe can obviouslythenonce theapplicationonce the artificial intelligence detects or predicts that there is an onset of acancerthis can be cross-checked by adoctorbut at least the initial screening processcan,FbxTVRfQFuI,00:04:42.800,"[{""timestamp"": ""00:04:42.800"", ""start_sec"": 282.8, ""end_sec"": 286.96, ""text"": ""will do because the training process"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:44.639"", ""start_sec"": 284.639, ""end_sec"": 289.04, ""text"": ""will be that much better and last but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:46.960"", ""start_sec"": 286.96, ""end_sec"": 291.6, ""text"": ""not least with deep learning you can"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:49.040"", ""start_sec"": 289.04, ""end_sec"": 293.36, ""text"": ""really avoid the manual process of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:51.600"", ""start_sec"": 291.6, ""end_sec"": 295.28, ""text"": ""feature extraction those are some of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:53.360"", ""start_sec"": 293.36, ""end_sec"": 297.12, ""text"": ""reasons why we need deep learning some"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:55.280"", ""start_sec"": 295.28, ""end_sec"": 300.16, ""text"": ""of the applications of deep learning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:04:57.120"", ""start_sec"": 297.12, ""end_sec"": 302.96, ""text"": ""deep learning has made major inroads and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:00.160"", ""start_sec"": 300.16, ""end_sec"": 305.44, ""text"": ""it is a major area in which deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:02.960"", ""start_sec"": 302.96, ""end_sec"": 308.88, ""text"": ""learning is applied is healthcare and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:05.440"", ""start_sec"": 305.44, ""end_sec"": 312.08, ""text"": ""within healthcare particularly oncology"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:08.880"", ""start_sec"": 308.88, ""end_sec"": 315.199, ""text"": ""which is basically cancer related stuff"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:12.080"", ""start_sec"": 312.08, ""end_sec"": 317.919, ""text"": ""one of the issues with cancer is that a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:15.199"", ""start_sec"": 315.199, ""end_sec"": 320.96, ""text"": ""lot of cancers today are curable they"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:17.919"", ""start_sec"": 317.919, ""end_sec"": 323.36, ""text"": ""can be cured they are detected early on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:20.960"", ""start_sec"": 320.96, ""end_sec"": 325.6, ""text"": ""and the challenge with that is when a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:23.360"", ""start_sec"": 323.36, ""end_sec"": 328.08, ""text"": ""diagnostics is performed let's say an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:25.600"", ""start_sec"": 325.6, ""end_sec"": 330.32, ""text"": ""image has been taken of a patient to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:28.080"", ""start_sec"": 328.08, ""end_sec"": 333.12, ""text"": ""detect whether there is cancer or not"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:30.320"", ""start_sec"": 330.32, ""end_sec"": 335.12, ""text"": ""you need a specialist to look at the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:33.120"", ""start_sec"": 333.12, ""end_sec"": 338.08, ""text"": ""image and determine whether it is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:35.120"", ""start_sec"": 335.12, ""end_sec"": 341.199, ""text"": ""patient is fine or there is any onset of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:38.080"", ""start_sec"": 338.08, ""end_sec"": 344.15999999999997, ""text"": ""cancer and the number of specialists are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:41.199"", ""start_sec"": 341.199, ""end_sec"": 346.639, ""text"": ""limited so if we use deep learning if we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:44.160"", ""start_sec"": 344.15999999999997, ""end_sec"": 348.88, ""text"": ""use automation here or if we use"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:46.639"", ""start_sec"": 346.639, ""end_sec"": 352.0, ""text"": ""artificial intelligence here then the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:48.880"", ""start_sec"": 348.88, ""end_sec"": 354.639, ""text"": ""system can with a certain amount of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:52.000"", ""start_sec"": 352.0, ""end_sec"": 357.52, ""text"": ""good amount of accuracy determine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:54.639"", ""start_sec"": 354.639, ""end_sec"": 360.0, ""text"": ""whether a particular patient is having"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:05:57.520"", ""start_sec"": 357.52, ""end_sec"": 362.96, ""text"": ""cancer or not so the prediction or the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:00.000"", ""start_sec"": 360.0, ""end_sec"": 365.919, ""text"": ""detection process of a disease like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:02.960"", ""start_sec"": 362.96, ""end_sec"": 368.16, ""text"": ""cancer can be expedited the detection"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:05.919"", ""start_sec"": 365.919, ""end_sec"": 370.8, ""text"": ""process can be expedited can be faster"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:08.160"", ""start_sec"": 368.16, ""end_sec"": 373.6, ""text"": ""without really waiting for a specialist"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:10.800"", ""start_sec"": 370.8, ""end_sec"": 375.919, ""text"": ""we can obviously then once the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:13.600"", ""start_sec"": 373.6, ""end_sec"": 378.479, ""text"": ""application once the artificial"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:15.919"", ""start_sec"": 375.919, ""end_sec"": 380.8, ""text"": ""intelligence detects or predicts that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:18.479"", ""start_sec"": 378.479, ""end_sec"": 383.12, ""text"": ""there is an onset of a cancer this can"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:20.800"", ""start_sec"": 380.8, ""end_sec"": 385.68, ""text"": ""be cross-checked by a doctor but at"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:23.120"", ""start_sec"": 383.12, ""end_sec"": 387.52, ""text"": ""least the initial screening process can"", ""video_id"": ""FbxTVRfQFuI""}]",282.8,387.52
beautomatedand that is where the current focus is with respect to deep learning inhealthcarewhatelserobotics is anotherareadeep learning is majorly used in robotics and you must haveseennowadays robots areeverywherehumanoidsthe industrial robots which are used for manufacturingprocessyou must have heard aboutsofiawho got citizenship with saudiarabiaand soonthere are multiple suchrobotswhich are knowledgeorientedbut there are also industrial robots are used inindustriesin the manufacturing processand increasinglyin security and also indefenseforexampleimageprocessingvideo is fed to them and they need to be able to detectobjectsobstacles and so on and soforthso that's where deep learning isusedthey need to be able to hear and make sense of the sounds that they arehearingthat needs deep learning aswellso robotics is a major area where deep learning isappliedthen we have self-driving cars or autonomouscarsyou must have heard of google's autonomouscarwhich has been tested for millions of miles and pretty much incidentfreetherewereofcoursea couple of incidents here andtherebut it is uh considered to be fairlysafeand there are today a lot of automotivecompaniesin fact pretty much every automotive company worth itsnameis investing in self-driving cars or autonomouscarsand it is predicted that in the next probably 10 to 15yearsthese will be in production and they will be used extensively in realliferight now they are all in rnd and in,FbxTVRfQFuI,00:06:25.680,"[{""timestamp"": ""00:06:25.680"", ""start_sec"": 385.68, ""end_sec"": 389.919, ""text"": ""be automated and that is where the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:27.520"", ""start_sec"": 387.52, ""end_sec"": 392.15999999999997, ""text"": ""current focus is with respect to deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:29.919"", ""start_sec"": 389.919, ""end_sec"": 394.56, ""text"": ""learning in healthcare what else"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:32.160"", ""start_sec"": 392.15999999999997, ""end_sec"": 398.319, ""text"": ""robotics is another area deep learning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:34.560"", ""start_sec"": 394.56, ""end_sec"": 400.88, ""text"": ""is majorly used in robotics and you must"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:38.319"", ""start_sec"": 398.319, ""end_sec"": 403.199, ""text"": ""have seen nowadays robots are everywhere"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:40.880"", ""start_sec"": 400.88, ""end_sec"": 405.12, ""text"": ""humanoids the industrial robots which"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:43.199"", ""start_sec"": 403.199, ""end_sec"": 408.08, ""text"": ""are used for manufacturing process you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:45.120"", ""start_sec"": 405.12, ""end_sec"": 410.639, ""text"": ""must have heard about sofia who got"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:48.080"", ""start_sec"": 408.08, ""end_sec"": 413.36, ""text"": ""citizenship with saudi arabia and so on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:50.639"", ""start_sec"": 410.639, ""end_sec"": 415.84000000000003, ""text"": ""there are multiple such robots which are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:53.360"", ""start_sec"": 413.36, ""end_sec"": 418.88, ""text"": ""knowledge oriented but there are also"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:55.840"", ""start_sec"": 415.84000000000003, ""end_sec"": 420.639, ""text"": ""industrial robots are used in industries"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:06:58.880"", ""start_sec"": 418.88, ""end_sec"": 423.12, ""text"": ""in the manufacturing process and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:00.639"", ""start_sec"": 420.639, ""end_sec"": 425.44, ""text"": ""increasingly in security and also in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:03.120"", ""start_sec"": 423.12, ""end_sec"": 427.84, ""text"": ""defense for example image processing"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:05.440"", ""start_sec"": 425.44, ""end_sec"": 430.08, ""text"": ""video is fed to them and they need to be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:07.840"", ""start_sec"": 427.84, ""end_sec"": 431.599, ""text"": ""able to detect objects obstacles and so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:10.080"", ""start_sec"": 430.08, ""end_sec"": 433.52, ""text"": ""on and so forth so that's where deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:11.599"", ""start_sec"": 431.599, ""end_sec"": 435.599, ""text"": ""learning is used they need to be able to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:13.520"", ""start_sec"": 433.52, ""end_sec"": 437.52, ""text"": ""hear and make sense of the sounds that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:15.599"", ""start_sec"": 435.599, ""end_sec"": 440.4, ""text"": ""they are hearing that needs deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:17.520"", ""start_sec"": 437.52, ""end_sec"": 442.8, ""text"": ""learning as well so robotics is a major"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:20.400"", ""start_sec"": 440.4, ""end_sec"": 445.68, ""text"": ""area where deep learning is applied then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:22.800"", ""start_sec"": 442.8, ""end_sec"": 447.919, ""text"": ""we have self-driving cars or autonomous"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:25.680"", ""start_sec"": 445.68, ""end_sec"": 450.96, ""text"": ""cars you must have heard of google's"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:27.919"", ""start_sec"": 447.919, ""end_sec"": 453.759, ""text"": ""autonomous car which has been tested for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:30.960"", ""start_sec"": 450.96, ""end_sec"": 455.44, ""text"": ""millions of miles and pretty much"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:33.759"", ""start_sec"": 453.759, ""end_sec"": 457.12, ""text"": ""incident free there were of course a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:35.440"", ""start_sec"": 455.44, ""end_sec"": 459.759, ""text"": ""couple of incidents here and there but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:37.120"", ""start_sec"": 457.12, ""end_sec"": 462.88, ""text"": ""it is uh considered to be fairly safe"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:39.759"", ""start_sec"": 459.759, ""end_sec"": 465.12, ""text"": ""and there are today a lot of automotive"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:42.880"", ""start_sec"": 462.88, ""end_sec"": 467.52, ""text"": ""companies in fact pretty much every"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:45.120"", ""start_sec"": 465.12, ""end_sec"": 469.919, ""text"": ""automotive company worth its name is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:47.520"", ""start_sec"": 467.52, ""end_sec"": 472.08, ""text"": ""investing in self-driving cars or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:49.919"", ""start_sec"": 469.919, ""end_sec"": 474.56, ""text"": ""autonomous cars and it is predicted that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:52.080"", ""start_sec"": 472.08, ""end_sec"": 476.15999999999997, ""text"": ""in the next probably 10 to 15 years"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:54.560"", ""start_sec"": 474.56, ""end_sec"": 479.12, ""text"": ""these will be in production and they"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:56.160"", ""start_sec"": 476.15999999999997, ""end_sec"": 481.039, ""text"": ""will be used extensively in real life"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:07:59.120"", ""start_sec"": 479.12, ""end_sec"": 483.039, ""text"": ""right now they are all in rnd and in"", ""video_id"": ""FbxTVRfQFuI""}]",385.68,483.039
testphasesbut pretty soon these will be on theroadso this is another area where deep learning isusedand how is itusedwhere is itusedwithin autonomousdrivingthe car actually is fed with video of surroundings and it is supposed to process thatinformationprocess that video and determine if there are anyobstaclesit has to determine if there are any cars in thesitewill detect whether it is driving in thelanealsoit has to determine whether the signal is green or red sothat accordinglyit can move forward orwaitso for all these videoanalysisdeep learning isusedin addition tothatthetrainingoverall training to drive thecarhappens in a deep learningenvironmentsoagaina lot of scope here to use deeplearninga couple of other applications are missiontranslationstoday we have a lot of information and very often this information is in one particularlanguageand more specifically inenglishand people need information in various parts of theworldit is pretty difficult for human beings to translate each and every piece of information or every document into all possiblelanguagesthere are probably at least hundreds of languages or if notmoreto translate each and every document into every language is prettydifficultthereforewe can use deep learning to do pretty much like a real-time translationmechanismso we don't have to translate everything and keep itreadybut we train applications or artificial intelligence systems that will do the translation on theflyfor,FbxTVRfQFuI,00:08:01.039,"[{""timestamp"": ""00:08:01.039"", ""start_sec"": 481.039, ""end_sec"": 485.36, ""text"": ""test phases but pretty soon these will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:03.039"", ""start_sec"": 483.039, ""end_sec"": 487.28, ""text"": ""be on the road so this is another area"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:05.360"", ""start_sec"": 485.36, ""end_sec"": 488.96, ""text"": ""where deep learning is used and how is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:07.280"", ""start_sec"": 487.28, ""end_sec"": 491.759, ""text"": ""it used where is it used within"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:08.960"", ""start_sec"": 488.96, ""end_sec"": 494.96, ""text"": ""autonomous driving the car actually is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:11.759"", ""start_sec"": 491.759, ""end_sec"": 497.039, ""text"": ""fed with video of surroundings and it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:14.960"", ""start_sec"": 494.96, ""end_sec"": 498.879, ""text"": ""supposed to process that information"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:17.039"", ""start_sec"": 497.039, ""end_sec"": 500.8, ""text"": ""process that video and determine if"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:18.879"", ""start_sec"": 498.879, ""end_sec"": 503.039, ""text"": ""there are any obstacles it has to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:20.800"", ""start_sec"": 500.8, ""end_sec"": 505.759, ""text"": ""determine if there are any cars in the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:23.039"", ""start_sec"": 503.039, ""end_sec"": 508.16, ""text"": ""site will detect whether it is driving"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:25.759"", ""start_sec"": 505.759, ""end_sec"": 511.759, ""text"": ""in the lane also it has to determine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:28.160"", ""start_sec"": 508.16, ""end_sec"": 514.159, ""text"": ""whether the signal is green or red so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:31.759"", ""start_sec"": 511.759, ""end_sec"": 517.76, ""text"": ""that accordingly it can move forward or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:34.159"", ""start_sec"": 514.159, ""end_sec"": 519.599, ""text"": ""wait so for all these video analysis"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:37.760"", ""start_sec"": 517.76, ""end_sec"": 521.919, ""text"": ""deep learning is used in addition to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:39.599"", ""start_sec"": 519.599, ""end_sec"": 524.72, ""text"": ""that the training overall training to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:41.919"", ""start_sec"": 521.919, ""end_sec"": 527.2, ""text"": ""drive the car happens in a deep learning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:44.720"", ""start_sec"": 524.72, ""end_sec"": 528.72, ""text"": ""environment so again a lot of scope here"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:47.200"", ""start_sec"": 527.2, ""end_sec"": 531.12, ""text"": ""to use deep learning a couple of other"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:48.720"", ""start_sec"": 528.72, ""end_sec"": 534.88, ""text"": ""applications are mission translations"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:51.120"", ""start_sec"": 531.12, ""end_sec"": 537.76, ""text"": ""today we have a lot of information and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:54.880"", ""start_sec"": 534.88, ""end_sec"": 539.519, ""text"": ""very often this information is in one"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:57.760"", ""start_sec"": 537.76, ""end_sec"": 543.12, ""text"": ""particular language and more"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:08:59.519"", ""start_sec"": 539.519, ""end_sec"": 545.519, ""text"": ""specifically in english and people need"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:03.120"", ""start_sec"": 543.12, ""end_sec"": 548.56, ""text"": ""information in various parts of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:05.519"", ""start_sec"": 545.519, ""end_sec"": 551.12, ""text"": ""world it is pretty difficult for human"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:08.560"", ""start_sec"": 548.56, ""end_sec"": 553.519, ""text"": ""beings to translate each and every piece"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:11.120"", ""start_sec"": 551.12, ""end_sec"": 555.279, ""text"": ""of information or every document into"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:13.519"", ""start_sec"": 553.519, ""end_sec"": 557.44, ""text"": ""all possible languages there are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:15.279"", ""start_sec"": 555.279, ""end_sec"": 559.6, ""text"": ""probably at least hundreds of languages"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:17.440"", ""start_sec"": 557.44, ""end_sec"": 562.72, ""text"": ""or if not more to translate each and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:19.600"", ""start_sec"": 559.6, ""end_sec"": 565.92, ""text"": ""every document into every language is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:22.720"", ""start_sec"": 562.72, ""end_sec"": 568.56, ""text"": ""pretty difficult therefore we can use"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:25.920"", ""start_sec"": 565.92, ""end_sec"": 571.44, ""text"": ""deep learning to do pretty much like a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:28.560"", ""start_sec"": 568.56, ""end_sec"": 573.2, ""text"": ""real-time translation mechanism so we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:31.440"", ""start_sec"": 571.44, ""end_sec"": 576.16, ""text"": ""don't have to translate everything and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:33.200"", ""start_sec"": 573.2, ""end_sec"": 578.64, ""text"": ""keep it ready but we train applications"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:36.160"", ""start_sec"": 576.16, ""end_sec"": 581.519, ""text"": ""or artificial intelligence systems that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:38.640"", ""start_sec"": 578.64, ""end_sec"": 584.56, ""text"": ""will do the translation on the fly for"", ""video_id"": ""FbxTVRfQFuI""}]",481.039,584.56
exampleyou go to somewhere like china and you want to know what is written on asignboardnow it is impossible for somebody to translate that and put it on the web or something likethatso you have an application which is trained to translate stuff on theflyso you probably this can be running on your mobilephoneon yoursmartphoneyou scanthisthe application will instantly translate that from chinese toenglishthat isonethen there could be web applications where there may be a research document which is all in maybe chinese orjapaneseand you want to translate that to study thatdocumentor in that case you need totranslatesothereforedeep learning is used in such situations aswelland that is again ondemandso it is not like you have to translate all these documents from other languages into english and one shot and keep itsomewherethatis againpretty much an impossibletaskbut on a neatbasisso you have systems that are trained to translate on theflyso mission translation is another major area where deep learning isusedthen there are a few other upcoming areas where synthesizing is done by neuralnetsforexamplemusic composition and generation ofmusicso you can train a neural net to producemusiceven to composemusicso this is a funthingthis is stillupcomingit needs a lot of effort to train such neuralnetit has been proved that it ispossibleso this is a relatively newareaand on the samelinescolorization ofimagesso these two images on the left hand side is a,FbxTVRfQFuI,00:09:41.519,"[{""timestamp"": ""00:09:41.519"", ""start_sec"": 581.519, ""end_sec"": 586.32, ""text"": ""example you go to somewhere like china"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:44.560"", ""start_sec"": 584.56, ""end_sec"": 588.48, ""text"": ""and you want to know what is written on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:46.320"", ""start_sec"": 586.32, ""end_sec"": 590.8, ""text"": ""a signboard now it is impossible for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:48.480"", ""start_sec"": 588.48, ""end_sec"": 592.8, ""text"": ""somebody to translate that and put it on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:50.800"", ""start_sec"": 590.8, ""end_sec"": 595.44, ""text"": ""the web or something like that so you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:52.800"", ""start_sec"": 592.8, ""end_sec"": 597.92, ""text"": ""have an application which is trained to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:55.440"", ""start_sec"": 595.44, ""end_sec"": 600.0, ""text"": ""translate stuff on the fly so you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:09:57.920"", ""start_sec"": 597.92, ""end_sec"": 602.24, ""text"": ""probably this can be running on your"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:00.000"", ""start_sec"": 600.0, ""end_sec"": 605.2, ""text"": ""mobile phone on your smartphone you scan"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:02.240"", ""start_sec"": 602.24, ""end_sec"": 607.44, ""text"": ""this the application will instantly"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:05.200"", ""start_sec"": 605.2, ""end_sec"": 610.24, ""text"": ""translate that from chinese to english"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:07.440"", ""start_sec"": 607.44, ""end_sec"": 611.76, ""text"": ""that is one then there could be web"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:10.240"", ""start_sec"": 610.24, ""end_sec"": 614.399, ""text"": ""applications where there may be a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:11.760"", ""start_sec"": 611.76, ""end_sec"": 616.64, ""text"": ""research document which is all in maybe"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:14.399"", ""start_sec"": 614.399, ""end_sec"": 619.8389999999999, ""text"": ""chinese or japanese and you want to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:16.640"", ""start_sec"": 616.64, ""end_sec"": 622.0, ""text"": ""translate that to study that document or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:19.839"", ""start_sec"": 619.8389999999999, ""end_sec"": 623.8389999999999, ""text"": ""in that case you need to translate so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:22.000"", ""start_sec"": 622.0, ""end_sec"": 626.16, ""text"": ""therefore deep learning is used in such"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:23.839"", ""start_sec"": 623.8389999999999, ""end_sec"": 628.16, ""text"": ""situations as well and that is again on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:26.160"", ""start_sec"": 626.16, ""end_sec"": 630.24, ""text"": ""demand so it is not like you have to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:28.160"", ""start_sec"": 628.16, ""end_sec"": 631.92, ""text"": ""translate all these documents from other"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:30.240"", ""start_sec"": 630.24, ""end_sec"": 634.0, ""text"": ""languages into english and one shot and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:31.920"", ""start_sec"": 631.92, ""end_sec"": 636.48, ""text"": ""keep it somewhere that is again pretty"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:34.000"", ""start_sec"": 634.0, ""end_sec"": 638.16, ""text"": ""much an impossible task but on a neat"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:36.480"", ""start_sec"": 636.48, ""end_sec"": 640.399, ""text"": ""basis so you have systems that are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:38.160"", ""start_sec"": 638.16, ""end_sec"": 642.0, ""text"": ""trained to translate on the fly so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:40.399"", ""start_sec"": 640.399, ""end_sec"": 643.6, ""text"": ""mission translation is another major"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:42.000"", ""start_sec"": 642.0, ""end_sec"": 645.92, ""text"": ""area where deep learning is used then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:43.600"", ""start_sec"": 643.6, ""end_sec"": 648.8, ""text"": ""there are a few other upcoming areas"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:45.920"", ""start_sec"": 645.92, ""end_sec"": 651.279, ""text"": ""where synthesizing is done by neural"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:48.800"", ""start_sec"": 648.8, ""end_sec"": 653.68, ""text"": ""nets for example music composition and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:51.279"", ""start_sec"": 651.279, ""end_sec"": 656.88, ""text"": ""generation of music so you can train a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:53.680"", ""start_sec"": 653.68, ""end_sec"": 659.68, ""text"": ""neural net to produce music even to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:56.880"", ""start_sec"": 656.88, ""end_sec"": 662.0, ""text"": ""compose music so this is a fun thing"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:10:59.680"", ""start_sec"": 659.68, ""end_sec"": 664.72, ""text"": ""this is still upcoming it needs a lot of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:02.000"", ""start_sec"": 662.0, ""end_sec"": 666.64, ""text"": ""effort to train such neural net it has"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:04.720"", ""start_sec"": 664.72, ""end_sec"": 669.12, ""text"": ""been proved that it is possible so this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:06.640"", ""start_sec"": 666.64, ""end_sec"": 671.76, ""text"": ""is a relatively new area and on the same"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:09.120"", ""start_sec"": 669.12, ""end_sec"": 673.92, ""text"": ""lines colorization of images so these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:11.760"", ""start_sec"": 671.76, ""end_sec"": 675.839, ""text"": ""two images on the left hand side is a"", ""video_id"": ""FbxTVRfQFuI""}]",581.519,675.839
grayscale image or a black and whiteimagethis was colored by a neural net or a deep learningapplicationas you canseeit's done a very good job of applying the colors and obviously this was trained to do thiscolorizationbutyesthis is one more application of deeplearningnowone of the major secret sauce of deep learning is neuralnetworkdeep learning works on neural network or consists of neuralnetworkso let us see what is neuralnetworkneural network or artificial neural network is designed or based on the humanbrainnowhuman brain consists of billions of small cells that are known asneuronsartificial neural networksisin awaytrying to simulate the humanbrainso this is a quick diagram of biologicalneurona biological neuron consists of the majorpartwhich is the cellnucleusand then it has sometentacleskind of stuff on the top calleddendriteand then there is like a long tailwhich is known as theaxonfurtheragainat the end of thisactionare what are known assynapsesthese in turn are connected to the dendrites of the nextneuronand all these neurons are interconnected with eachotherthereforethey are like billions of them sitting in our brain and they're allactivethey'reworkingthey based on thesignalsthey receive signals as inputs from other neurons or maybe from other parts of the bodyandbased on certaincriteriathey send signals to the neurons at the otherendso they they get either activated or they don't getactivatedbasedonso it is like a binarygatesso they get,FbxTVRfQFuI,00:11:13.920,"[{""timestamp"": ""00:11:13.920"", ""start_sec"": 673.92, ""end_sec"": 678.72, ""text"": ""grayscale image or a black and white"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:15.839"", ""start_sec"": 675.839, ""end_sec"": 680.48, ""text"": ""image this was colored by a neural net"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:18.720"", ""start_sec"": 678.72, ""end_sec"": 682.959, ""text"": ""or a deep learning application as you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:20.480"", ""start_sec"": 680.48, ""end_sec"": 685.04, ""text"": ""can see it's done a very good job of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:22.959"", ""start_sec"": 682.959, ""end_sec"": 688.0, ""text"": ""applying the colors and obviously this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:25.040"", ""start_sec"": 685.04, ""end_sec"": 690.32, ""text"": ""was trained to do this colorization but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:28.000"", ""start_sec"": 688.0, ""end_sec"": 693.36, ""text"": ""yes this is one more application of deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:30.320"", ""start_sec"": 690.32, ""end_sec"": 697.279, ""text"": ""learning now one of the major secret"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:33.360"", ""start_sec"": 693.36, ""end_sec"": 700.16, ""text"": ""sauce of deep learning is neural network"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:37.279"", ""start_sec"": 697.279, ""end_sec"": 702.24, ""text"": ""deep learning works on neural network or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:40.160"", ""start_sec"": 700.16, ""end_sec"": 705.279, ""text"": ""consists of neural network so let us see"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:42.240"", ""start_sec"": 702.24, ""end_sec"": 709.04, ""text"": ""what is neural network neural network or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:45.279"", ""start_sec"": 705.279, ""end_sec"": 713.36, ""text"": ""artificial neural network is designed or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:49.040"", ""start_sec"": 709.04, ""end_sec"": 716.88, ""text"": ""based on the human brain now human brain"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:53.360"", ""start_sec"": 713.36, ""end_sec"": 719.519, ""text"": ""consists of billions of small cells that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:56.880"", ""start_sec"": 716.88, ""end_sec"": 723.12, ""text"": ""are known as neurons artificial neural"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:11:59.519"", ""start_sec"": 719.519, ""end_sec"": 725.519, ""text"": ""networks is in a way trying to simulate"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:03.120"", ""start_sec"": 723.12, ""end_sec"": 727.839, ""text"": ""the human brain so this is a quick"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:05.519"", ""start_sec"": 725.519, ""end_sec"": 730.399, ""text"": ""diagram of biological neuron a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:07.839"", ""start_sec"": 727.839, ""end_sec"": 732.959, ""text"": ""biological neuron consists of the major"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:10.399"", ""start_sec"": 730.399, ""end_sec"": 736.079, ""text"": ""part which is the cell nucleus and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:12.959"", ""start_sec"": 732.959, ""end_sec"": 738.24, ""text"": ""it has some tentacles kind of stuff on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:16.079"", ""start_sec"": 736.079, ""end_sec"": 740.16, ""text"": ""the top called dendrite and then there"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:18.240"", ""start_sec"": 738.24, ""end_sec"": 742.399, ""text"": ""is like a long tail which is known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:20.160"", ""start_sec"": 740.16, ""end_sec"": 744.24, ""text"": ""the axon further again at the end of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:22.399"", ""start_sec"": 742.399, ""end_sec"": 747.68, ""text"": ""this action are what are known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:24.240"", ""start_sec"": 744.24, ""end_sec"": 750.88, ""text"": ""synapses these in turn are connected to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:27.680"", ""start_sec"": 747.68, ""end_sec"": 753.68, ""text"": ""the dendrites of the next neuron and all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:30.880"", ""start_sec"": 750.88, ""end_sec"": 755.44, ""text"": ""these neurons are interconnected with"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:33.680"", ""start_sec"": 753.68, ""end_sec"": 757.519, ""text"": ""each other therefore they are like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:35.440"", ""start_sec"": 755.44, ""end_sec"": 759.44, ""text"": ""billions of them sitting in our brain"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:37.519"", ""start_sec"": 757.519, ""end_sec"": 762.0, ""text"": ""and they're all active they're working"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:39.440"", ""start_sec"": 759.44, ""end_sec"": 765.36, ""text"": ""they based on the signals they receive"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:42.000"", ""start_sec"": 762.0, ""end_sec"": 767.92, ""text"": ""signals as inputs from other neurons or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:45.360"", ""start_sec"": 765.36, ""end_sec"": 770.639, ""text"": ""maybe from other parts of the body and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:47.920"", ""start_sec"": 767.92, ""end_sec"": 772.72, ""text"": ""based on certain criteria they send"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:50.639"", ""start_sec"": 770.639, ""end_sec"": 774.8, ""text"": ""signals to the neurons at the other end"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:52.720"", ""start_sec"": 772.72, ""end_sec"": 776.88, ""text"": ""so they they get either activated or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:54.800"", ""start_sec"": 774.8, ""end_sec"": 779.76, ""text"": ""they don't get activated based on so it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:12:56.880"", ""start_sec"": 776.88, ""end_sec"": 782.48, ""text"": ""is like a binary gates so they get"", ""video_id"": ""FbxTVRfQFuI""}]",673.92,782.48
activated or notactivatedbased on the inputs that they receive and soonso we will see a little bit of those details as we move forward in our artificialneuronbut this is a biologicalneuronthis is the structure of a biological neuron and artificial neural network is based on the humanbrainthe smallest component of artificial neural network is an artificialneuronas shownheresometimes is also referred to asperceptronnow this is a very high leveldiagramthe artificial neuron has a small central unit which will receive the input if it isdoinglet'ssayimageprocessingthe inputs could be pixel values of theimagewhich is represented here asx1x2 and sooneach of the inputs are multiplied by what is known asweightswhich are represented asw1w2 and soonthere is in the centralunitbasically there is a summation of these weighted inputs which is like x1 into w1 plus x2 into w2 and soonthe products are then added and then there is a bias that is added tothatin the next slide we willseethat passes through an activation function and the output comes as aywhich is theoutputand based on certaincriteriathe cell gets either activated or notactivatedso this output would be like a zero or a one binaryformatokayso we will see that in a little bit moredetailbut let's do a quick comparison between biological and artificialneuronsjust like a biologicalneuronthere are dendrites and then there is a cell nucleus and synapse and anaxonwe have in the artificial neuron aswell,FbxTVRfQFuI,00:12:59.760,"[{""timestamp"": ""00:12:59.760"", ""start_sec"": 779.76, ""end_sec"": 784.8, ""text"": ""activated or not activated based on the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:02.480"", ""start_sec"": 782.48, ""end_sec"": 786.399, ""text"": ""inputs that they receive and so on so we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:04.800"", ""start_sec"": 784.8, ""end_sec"": 788.72, ""text"": ""will see a little bit of those details"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:06.399"", ""start_sec"": 786.399, ""end_sec"": 790.88, ""text"": ""as we move forward in our artificial"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:08.720"", ""start_sec"": 788.72, ""end_sec"": 792.32, ""text"": ""neuron but this is a biological neuron"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:10.880"", ""start_sec"": 790.88, ""end_sec"": 795.2, ""text"": ""this is the structure of a biological"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:12.320"", ""start_sec"": 792.32, ""end_sec"": 797.68, ""text"": ""neuron and artificial neural network is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:15.200"", ""start_sec"": 795.2, ""end_sec"": 800.32, ""text"": ""based on the human brain the smallest"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:17.680"", ""start_sec"": 797.68, ""end_sec"": 803.44, ""text"": ""component of artificial neural network"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:20.320"", ""start_sec"": 800.32, ""end_sec"": 805.8389999999999, ""text"": ""is an artificial neuron as shown here"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:23.440"", ""start_sec"": 803.44, ""end_sec"": 808.0, ""text"": ""sometimes is also referred to as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:25.839"", ""start_sec"": 805.8389999999999, ""end_sec"": 810.24, ""text"": ""perceptron now this is a very high level"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:28.000"", ""start_sec"": 808.0, ""end_sec"": 812.8, ""text"": ""diagram the artificial neuron has a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:30.240"", ""start_sec"": 810.24, ""end_sec"": 815.76, ""text"": ""small central unit which will receive"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:32.800"", ""start_sec"": 812.8, ""end_sec"": 818.32, ""text"": ""the input if it is doing let's say image"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:35.760"", ""start_sec"": 815.76, ""end_sec"": 821.04, ""text"": ""processing the inputs could be pixel"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:38.320"", ""start_sec"": 818.32, ""end_sec"": 824.48, ""text"": ""values of the image which is represented"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:41.040"", ""start_sec"": 821.04, ""end_sec"": 827.68, ""text"": ""here as x1 x2 and so on each of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:44.480"", ""start_sec"": 824.48, ""end_sec"": 830.32, ""text"": ""inputs are multiplied by what is known"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:47.680"", ""start_sec"": 827.68, ""end_sec"": 833.2, ""text"": ""as weights which are represented as w1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:50.320"", ""start_sec"": 830.32, ""end_sec"": 836.24, ""text"": ""w2 and so on there is in the central"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:53.200"", ""start_sec"": 833.2, ""end_sec"": 839.6, ""text"": ""unit basically there is a summation of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:56.240"", ""start_sec"": 836.24, ""end_sec"": 843.279, ""text"": ""these weighted inputs which is like x1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:13:59.600"", ""start_sec"": 839.6, ""end_sec"": 846.16, ""text"": ""into w1 plus x2 into w2 and so on the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:03.279"", ""start_sec"": 843.279, ""end_sec"": 848.079, ""text"": ""products are then added and then there"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:06.160"", ""start_sec"": 846.16, ""end_sec"": 850.72, ""text"": ""is a bias that is added to that in the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:08.079"", ""start_sec"": 848.079, ""end_sec"": 852.959, ""text"": ""next slide we will see that passes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:10.720"", ""start_sec"": 850.72, ""end_sec"": 856.16, ""text"": ""through an activation function and the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:12.959"", ""start_sec"": 852.959, ""end_sec"": 858.72, ""text"": ""output comes as a y which is the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:16.160"", ""start_sec"": 856.16, ""end_sec"": 860.88, ""text"": ""and based on certain criteria the cell"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:18.720"", ""start_sec"": 858.72, ""end_sec"": 863.519, ""text"": ""gets either activated or not activated"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:20.880"", ""start_sec"": 860.88, ""end_sec"": 866.959, ""text"": ""so this output would be like a zero or a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:23.519"", ""start_sec"": 863.519, ""end_sec"": 868.639, ""text"": ""one binary format okay so we will see"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:26.959"", ""start_sec"": 866.959, ""end_sec"": 870.639, ""text"": ""that in a little bit more detail but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:28.639"", ""start_sec"": 868.639, ""end_sec"": 873.04, ""text"": ""let's do a quick comparison between"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:30.639"", ""start_sec"": 870.639, ""end_sec"": 875.04, ""text"": ""biological and artificial neurons just"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:33.040"", ""start_sec"": 873.04, ""end_sec"": 876.639, ""text"": ""like a biological neuron there are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:35.040"", ""start_sec"": 875.04, ""end_sec"": 879.6, ""text"": ""dendrites and then there is a cell"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:36.639"", ""start_sec"": 876.639, ""end_sec"": 882.88, ""text"": ""nucleus and synapse and an axon"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:39.600"", ""start_sec"": 879.6, ""end_sec"": 885.92, ""text"": ""we have in the artificial neuron as well"", ""video_id"": ""FbxTVRfQFuI""}]",779.76,885.92
these inputs come like the deadrightif you will act like thedendritesthere is a like a central unit which performs the summation ofthese uhweightedinputswhich is basicallyw1 x1 w2x2 and soonand then our bias is added here and then that passes through what is known as an activationfunctionokayso these are known as theweightsw1w2and then there is a bias which will come out here and that isaddedthe biasisby thewaycommon for a particularneuronso there won't be likeb1 b2b3 and soononly weights will be one perinputthe bias is common for the entireneuronit is also commonforor the value of the bias remains the sameforall the neurons in a particularlayerwe will also see this as we move forward and we see deep neural network where there are multipleneuronsso that's theoutputnow the whole exercise of training the neuron is about changing these weights andbiasesas imentionedartificial neural network will consist of several such neurons and as a part of the trainingprocessthese weights keepchanginginitially they are assigned some random values through the trainingprocesstheweightsthe whole process of training is to come up with the optimum values ofw1w2 and wn and then the b4 or the bias for this particularneuronsuch that it gives an accurate output asrequiredso let's see what exactly thatmeansso the trainingprocessthis is how ithappensit takes theinputseach input is multiplied by a weight and theseweightsduringtrainingkeepchangingso initially they are assigned some random values and based on,FbxTVRfQFuI,00:14:42.880,"[{""timestamp"": ""00:14:42.880"", ""start_sec"": 882.88, ""end_sec"": 888.16, ""text"": ""these inputs come like the dead right if"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:45.920"", ""start_sec"": 885.92, ""end_sec"": 890.32, ""text"": ""you will act like the dendrites there is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:48.160"", ""start_sec"": 888.16, ""end_sec"": 892.88, ""text"": ""a like a central unit which performs the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:50.320"", ""start_sec"": 890.32, ""end_sec"": 896.16, ""text"": ""summation of these uh weighted inputs"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:52.880"", ""start_sec"": 892.88, ""end_sec"": 898.88, ""text"": ""which is basically w1 x1 w2 x2 and so on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:56.160"", ""start_sec"": 896.16, ""end_sec"": 900.639, ""text"": ""and then our bias is added here and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:14:58.880"", ""start_sec"": 898.88, ""end_sec"": 902.88, ""text"": ""that passes through what is known as an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:00.639"", ""start_sec"": 900.639, ""end_sec"": 904.639, ""text"": ""activation function okay so these are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:02.880"", ""start_sec"": 902.88, ""end_sec"": 906.88, ""text"": ""known as the weights w1 w2 and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:04.639"", ""start_sec"": 904.639, ""end_sec"": 909.519, ""text"": ""there is a bias which will come out here"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:06.880"", ""start_sec"": 906.88, ""end_sec"": 911.6, ""text"": ""and that is added the bias is by the way"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:09.519"", ""start_sec"": 909.519, ""end_sec"": 914.32, ""text"": ""common for a particular neuron so there"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:11.600"", ""start_sec"": 911.6, ""end_sec"": 916.8, ""text"": ""won't be like b1 b2 b3 and so on only"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:14.320"", ""start_sec"": 914.32, ""end_sec"": 919.44, ""text"": ""weights will be one per input the bias"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:16.800"", ""start_sec"": 916.8, ""end_sec"": 922.639, ""text"": ""is common for the entire neuron it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:19.440"", ""start_sec"": 919.44, ""end_sec"": 925.36, ""text"": ""also common for or the value of the bias"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:22.639"", ""start_sec"": 922.639, ""end_sec"": 928.0, ""text"": ""remains the same for all the neurons in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:25.360"", ""start_sec"": 925.36, ""end_sec"": 929.92, ""text"": ""a particular layer we will also see this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:28.000"", ""start_sec"": 928.0, ""end_sec"": 931.6, ""text"": ""as we move forward and we see deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:29.920"", ""start_sec"": 929.92, ""end_sec"": 934.16, ""text"": ""neural network where there are multiple"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:31.600"", ""start_sec"": 931.6, ""end_sec"": 937.92, ""text"": ""neurons so that's the output now the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:34.160"", ""start_sec"": 934.16, ""end_sec"": 941.519, ""text"": ""whole exercise of training the neuron is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:37.920"", ""start_sec"": 937.92, ""end_sec"": 943.519, ""text"": ""about changing these weights and biases"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:41.519"", ""start_sec"": 941.519, ""end_sec"": 946.0, ""text"": ""as i mentioned artificial neural network"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:43.519"", ""start_sec"": 943.519, ""end_sec"": 948.56, ""text"": ""will consist of several such neurons and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:46.000"", ""start_sec"": 946.0, ""end_sec"": 950.88, ""text"": ""as a part of the training process these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:48.560"", ""start_sec"": 948.56, ""end_sec"": 953.12, ""text"": ""weights keep changing initially they are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:50.880"", ""start_sec"": 950.88, ""end_sec"": 955.36, ""text"": ""assigned some random values through the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:53.120"", ""start_sec"": 953.12, ""end_sec"": 957.279, ""text"": ""training process the weights the whole"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:55.360"", ""start_sec"": 955.36, ""end_sec"": 960.88, ""text"": ""process of training is to come up with"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:15:57.279"", ""start_sec"": 957.279, ""end_sec"": 962.959, ""text"": ""the optimum values of w1 w2 and wn and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:00.880"", ""start_sec"": 960.88, ""end_sec"": 965.519, ""text"": ""then the b4 or the bias for this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:02.959"", ""start_sec"": 962.959, ""end_sec"": 968.399, ""text"": ""particular neuron such that it gives an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:05.519"", ""start_sec"": 965.519, ""end_sec"": 971.04, ""text"": ""accurate output as required so let's see"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:08.399"", ""start_sec"": 968.399, ""end_sec"": 973.44, ""text"": ""what exactly that means so the training"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:11.040"", ""start_sec"": 971.04, ""end_sec"": 976.72, ""text"": ""process this is how it happens it takes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:13.440"", ""start_sec"": 973.44, ""end_sec"": 979.04, ""text"": ""the inputs each input is multiplied by a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:16.720"", ""start_sec"": 976.72, ""end_sec"": 980.639, ""text"": ""weight and these weights during training"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:19.040"", ""start_sec"": 979.04, ""end_sec"": 983.44, ""text"": ""keep changing so initially they are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:20.639"", ""start_sec"": 980.639, ""end_sec"": 985.519, ""text"": ""assigned some random values and based on"", ""video_id"": ""FbxTVRfQFuI""}]",882.88,985.519
theoutputwhether it is correct orwrongthere is a feedback coming back and that will basically change these weights until it starts giving the rightoutputthat is represented in here as sigmaigoing from 1 to n if there are ninputs wiintoxiso this is the product ofw1 x1 w2x2 and soon rightand there is a bias that gets added here and that entire thing goes to what is known as an activationfunctionso essentially this is sigma of wi x iplus a value ofbiaswhich isabso that entire thing goes as an input to an activationfunctionnow this activation function takes this as aninputgives the output as a binaryoutputit could be a zero or aonethereareofcourseto startwithlet's assume it's a binaryoutputlater we will see that there are different types of activationfunctionsso it need not always be binaryoutputbut to start with let's keepsimpleso it decides whether the neuron should be fired ornotso that is theoutputlike a binaryoutput0 or1.allrightso again let me summarizethisso it takes theinputsso if you're processing animageforexamplethe inputs are the pixel values of theimagex1 x2up toxnthere could be hundreds oftheseso all of those are fedasso these are somevaluesand these pixel values again can be from 0 to56each of those pixel values are then multiplied with what is known as aweightthis is a numericvaluecan be anyvalueso this is anumberw1similarlyw2 is anumberso initially some random values will be assigned and each of these weights are,FbxTVRfQFuI,00:16:23.440,"[{""timestamp"": ""00:16:23.440"", ""start_sec"": 983.44, ""end_sec"": 987.519, ""text"": ""the output whether it is correct or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:25.519"", ""start_sec"": 985.519, ""end_sec"": 989.759, ""text"": ""wrong there is a feedback coming back"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:27.519"", ""start_sec"": 987.519, ""end_sec"": 993.12, ""text"": ""and that will basically change these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:29.759"", ""start_sec"": 989.759, ""end_sec"": 996.32, ""text"": ""weights until it starts giving the right"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:33.120"", ""start_sec"": 993.12, ""end_sec"": 999.199, ""text"": ""output that is represented in here as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:36.320"", ""start_sec"": 996.32, ""end_sec"": 1002.32, ""text"": ""sigma i going from 1 to n if there are n"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:39.199"", ""start_sec"": 999.199, ""end_sec"": 1006.16, ""text"": ""inputs wi into x i so this is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:42.320"", ""start_sec"": 1002.32, ""end_sec"": 1009.92, ""text"": ""product of w1 x1 w2 x2 and so on right"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:46.160"", ""start_sec"": 1006.16, ""end_sec"": 1012.9590000000001, ""text"": ""and there is a bias that gets added here"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:49.920"", ""start_sec"": 1009.92, ""end_sec"": 1015.36, ""text"": ""and that entire thing goes to what is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:52.959"", ""start_sec"": 1012.9590000000001, ""end_sec"": 1019.12, ""text"": ""known as an activation function so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:55.360"", ""start_sec"": 1015.36, ""end_sec"": 1022.16, ""text"": ""essentially this is sigma of w i x i"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:16:59.120"", ""start_sec"": 1019.12, ""end_sec"": 1025.36, ""text"": ""plus a value of bias which is a b so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:02.160"", ""start_sec"": 1022.16, ""end_sec"": 1027.919, ""text"": ""that entire thing goes as an input to an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:05.360"", ""start_sec"": 1025.36, ""end_sec"": 1030.48, ""text"": ""activation function now this activation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:07.919"", ""start_sec"": 1027.919, ""end_sec"": 1033.52, ""text"": ""function takes this as an input gives"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:10.480"", ""start_sec"": 1030.48, ""end_sec"": 1035.439, ""text"": ""the output as a binary output it could"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:13.520"", ""start_sec"": 1033.52, ""end_sec"": 1037.439, ""text"": ""be a zero or a one there are of course"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:15.439"", ""start_sec"": 1035.439, ""end_sec"": 1038.959, ""text"": ""to start with let's assume it's a binary"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:17.439"", ""start_sec"": 1037.439, ""end_sec"": 1040.799, ""text"": ""output later we will see that there are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:18.959"", ""start_sec"": 1038.959, ""end_sec"": 1043.12, ""text"": ""different types of activation functions"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:20.799"", ""start_sec"": 1040.799, ""end_sec"": 1045.439, ""text"": ""so it need not always be binary output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:23.120"", ""start_sec"": 1043.12, ""end_sec"": 1048.16, ""text"": ""but to start with let's keep simple so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:25.439"", ""start_sec"": 1045.439, ""end_sec"": 1050.799, ""text"": ""it decides whether the neuron should be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:28.160"", ""start_sec"": 1048.16, ""end_sec"": 1053.28, ""text"": ""fired or not so that is the output like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:30.799"", ""start_sec"": 1050.799, ""end_sec"": 1055.28, ""text"": ""a binary output 0 or 1. all right so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:33.280"", ""start_sec"": 1053.28, ""end_sec"": 1056.96, ""text"": ""again let me summarize this so it takes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:35.280"", ""start_sec"": 1055.28, ""end_sec"": 1059.28, ""text"": ""the inputs so if you're processing an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:36.960"", ""start_sec"": 1056.96, ""end_sec"": 1062.559, ""text"": ""image for example the inputs are the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:39.280"", ""start_sec"": 1059.28, ""end_sec"": 1064.559, ""text"": ""pixel values of the image x1 x2 up to xn"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:42.559"", ""start_sec"": 1062.559, ""end_sec"": 1066.48, ""text"": ""there could be hundreds of these so all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:44.559"", ""start_sec"": 1064.559, ""end_sec"": 1068.559, ""text"": ""of those are fed as so these are some"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:46.480"", ""start_sec"": 1066.48, ""end_sec"": 1071.2, ""text"": ""values and these pixel values again can"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:48.559"", ""start_sec"": 1068.559, ""end_sec"": 1074.4, ""text"": ""be from 0 to 56 each of those pixel"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:51.200"", ""start_sec"": 1071.2, ""end_sec"": 1076.16, ""text"": ""values are then multiplied with what is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:54.400"", ""start_sec"": 1074.4, ""end_sec"": 1078.16, ""text"": ""known as a weight this is a numeric"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:56.160"", ""start_sec"": 1076.16, ""end_sec"": 1081.36, ""text"": ""value can be any value so this is a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:17:58.160"", ""start_sec"": 1078.16, ""end_sec"": 1083.679, ""text"": ""number w1 similarly w2 is a number so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:01.360"", ""start_sec"": 1081.36, ""end_sec"": 1085.6, ""text"": ""initially some random values will be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:03.679"", ""start_sec"": 1083.679, ""end_sec"": 1087.52, ""text"": ""assigned and each of these weights are"", ""video_id"": ""FbxTVRfQFuI""}]",983.44,1087.52
multiplied with the input value and theirsumthis is known as the weightedsumso that is performed in this kind of the central unit and then a bias isaddedremember the bias is common for eachneuronso this is not the biasvalueis not one bias value for perinputso just keep that inmindthe biasvaluethere is one bias perneuronso it is likethissummation plus bias is the output from thesectionthis is not the complete output of theneuronbut this is the bias for output for steponethat goes as an input to what is known as activationfunctionand that activation function results in anoutputusually a binaryoutputlike a zero or a one which is known as the firing of theneuronokay goodso we talked about activationfunctionso what is an activationfunctionan activation function basically takes the weightedsumwhichiswesaw w1 x1 w2x2the sum of all that plus thebiasso it takes that as an input and it generates a certainoutputnowthere are different types of activation functions and the output is different for different types of activationfunctionsmoreoverwhy is an activation functionrequiredit is basically required to bring innon-linearitythat's the main reason why an activation function isrequiredso what are the different types of activationfunctionsthere are several types of activationfunctionsbut these are the most commononesthese are the ones that are currently inusesigmoid function was one of the earlyactivation,FbxTVRfQFuI,00:18:05.600,"[{""timestamp"": ""00:18:05.600"", ""start_sec"": 1085.6, ""end_sec"": 1089.919, ""text"": ""multiplied with the input value and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:07.520"", ""start_sec"": 1087.52, ""end_sec"": 1092.32, ""text"": ""their sum this is known as the weighted"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:09.919"", ""start_sec"": 1089.919, ""end_sec"": 1094.96, ""text"": ""sum so that is performed in this kind of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:12.320"", ""start_sec"": 1092.32, ""end_sec"": 1097.44, ""text"": ""the central unit and then a bias is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:14.960"", ""start_sec"": 1094.96, ""end_sec"": 1100.08, ""text"": ""added remember the bias is common for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:17.440"", ""start_sec"": 1097.44, ""end_sec"": 1101.76, ""text"": ""each neuron so this is not the bias"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:20.080"", ""start_sec"": 1100.08, ""end_sec"": 1104.559, ""text"": ""value is not one"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:21.760"", ""start_sec"": 1101.76, ""end_sec"": 1106.64, ""text"": ""bias value for per input so just keep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:24.559"", ""start_sec"": 1104.559, ""end_sec"": 1108.64, ""text"": ""that in mind the bias value there is one"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:26.640"", ""start_sec"": 1106.64, ""end_sec"": 1111.36, ""text"": ""bias per neuron so it is like this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:28.640"", ""start_sec"": 1108.64, ""end_sec"": 1113.2, ""text"": ""summation plus bias is the output from"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:31.360"", ""start_sec"": 1111.36, ""end_sec"": 1114.88, ""text"": ""the section this is not the complete"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:33.200"", ""start_sec"": 1113.2, ""end_sec"": 1117.6, ""text"": ""output of the neuron but this is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:34.880"", ""start_sec"": 1114.88, ""end_sec"": 1119.2, ""text"": ""bias for output for step one that goes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:37.600"", ""start_sec"": 1117.6, ""end_sec"": 1121.52, ""text"": ""as an input to what is known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:39.200"", ""start_sec"": 1119.2, ""end_sec"": 1124.32, ""text"": ""activation function and that activation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:41.520"", ""start_sec"": 1121.52, ""end_sec"": 1126.72, ""text"": ""function results in an output usually a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:44.320"", ""start_sec"": 1124.32, ""end_sec"": 1129.44, ""text"": ""binary output like a zero or a one which"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:46.720"", ""start_sec"": 1126.72, ""end_sec"": 1131.919, ""text"": ""is known as the firing of the neuron"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:49.440"", ""start_sec"": 1129.44, ""end_sec"": 1133.84, ""text"": ""okay good so we talked about activation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:51.919"", ""start_sec"": 1131.919, ""end_sec"": 1135.76, ""text"": ""function so what is an activation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:53.840"", ""start_sec"": 1133.84, ""end_sec"": 1138.88, ""text"": ""function an activation function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:55.760"", ""start_sec"": 1135.76, ""end_sec"": 1142.64, ""text"": ""basically takes the weighted sum which"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:18:58.880"", ""start_sec"": 1138.88, ""end_sec"": 1145.52, ""text"": ""is we saw w1 x1 w2 x2 the sum of all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:02.640"", ""start_sec"": 1142.64, ""end_sec"": 1148.799, ""text"": ""that plus the bias so it takes that as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:05.520"", ""start_sec"": 1145.52, ""end_sec"": 1150.64, ""text"": ""an input and it generates a certain"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:08.799"", ""start_sec"": 1148.799, ""end_sec"": 1152.64, ""text"": ""output now there are different types of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:10.640"", ""start_sec"": 1150.64, ""end_sec"": 1154.16, ""text"": ""activation functions and the output is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:12.640"", ""start_sec"": 1152.64, ""end_sec"": 1156.72, ""text"": ""different for different types of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:14.160"", ""start_sec"": 1154.16, ""end_sec"": 1158.72, ""text"": ""activation functions moreover why is an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:16.720"", ""start_sec"": 1156.72, ""end_sec"": 1160.96, ""text"": ""activation function required it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:18.720"", ""start_sec"": 1158.72, ""end_sec"": 1163.52, ""text"": ""basically required to bring in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:20.960"", ""start_sec"": 1160.96, ""end_sec"": 1165.76, ""text"": ""non-linearity that's the main reason why"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:23.520"", ""start_sec"": 1163.52, ""end_sec"": 1166.88, ""text"": ""an activation function is required so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:25.760"", ""start_sec"": 1165.76, ""end_sec"": 1168.72, ""text"": ""what are the different types of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:26.880"", ""start_sec"": 1166.88, ""end_sec"": 1170.72, ""text"": ""activation functions there are several"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:28.720"", ""start_sec"": 1168.72, ""end_sec"": 1172.72, ""text"": ""types of activation functions but these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:30.720"", ""start_sec"": 1170.72, ""end_sec"": 1175.2, ""text"": ""are the most common ones these are the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:32.720"", ""start_sec"": 1172.72, ""end_sec"": 1177.6, ""text"": ""ones that are currently in use sigmoid"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:35.200"", ""start_sec"": 1175.2, ""end_sec"": 1181.44, ""text"": ""function was one of the early activation"", ""video_id"": ""FbxTVRfQFuI""}]",1085.6,1181.44
functionsbut today relu has kind of takenoverso relu is by far the most popular activation function that is usedtodaybut still sigmoid function is still used in manysituationsthese different types of activation functions are used in different situations based on the kind of problem we are trying tosolveso what exactly is the difference between thesetwosigmoid gives the values of the output will be between 0 and1.threshold function is the value will be 0 up to a certain value and beyond that this is also known as a stepfunctionand beyond that it will be1.in case of sigmoid there is a gradualincreasebut in case of threshold it's like also known as a stepfunctionthere's a rapid or instantaneous change from zero toonewhereas insigmoidwe will see in the nextslidethere is a gradualincreasebut the value in this case is between zero and one aswellnow relufunctionon the otherhandit is equaltobasicallyif the input is 0 or less than0then the output is0whereas if the input is greater than0then the output is equal to theinputi know it's a littleconfusingbut in the next slides where we show the relu function it will becomeclearsimilarlyhyperbolictangentthis is similar to sigmoid in terms of the shape of thefunctionhoweverwhile sigmoid goes from 0 to1hyperbolic tangent goes from -1 to 1 and hereagainthe increase or the change from -1 to 1 is gradual and not like threshold or stepfunctionwhere it happensinstantaneouslyso let's take a,FbxTVRfQFuI,00:19:37.600,"[{""timestamp"": ""00:19:37.600"", ""start_sec"": 1177.6, ""end_sec"": 1184.4, ""text"": ""functions but today relu has kind of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:41.440"", ""start_sec"": 1181.44, ""end_sec"": 1186.96, ""text"": ""taken over so relu is by far the most"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:44.400"", ""start_sec"": 1184.4, ""end_sec"": 1189.6, ""text"": ""popular activation function that is used"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:46.960"", ""start_sec"": 1186.96, ""end_sec"": 1192.32, ""text"": ""today but still sigmoid function is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:49.600"", ""start_sec"": 1189.6, ""end_sec"": 1194.16, ""text"": ""still used in many situations these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:52.320"", ""start_sec"": 1192.32, ""end_sec"": 1196.4, ""text"": ""different types of activation functions"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:54.160"", ""start_sec"": 1194.16, ""end_sec"": 1198.08, ""text"": ""are used in different situations based"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:56.400"", ""start_sec"": 1196.4, ""end_sec"": 1200.0, ""text"": ""on the kind of problem we are trying to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:19:58.080"", ""start_sec"": 1198.08, ""end_sec"": 1201.84, ""text"": ""solve so what exactly is the difference"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:00.000"", ""start_sec"": 1200.0, ""end_sec"": 1203.919, ""text"": ""between these two sigmoid gives the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:01.840"", ""start_sec"": 1201.84, ""end_sec"": 1206.799, ""text"": ""values of the output will be between 0"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:03.919"", ""start_sec"": 1203.919, ""end_sec"": 1207.76, ""text"": ""and 1. threshold function is the value"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:06.799"", ""start_sec"": 1206.799, ""end_sec"": 1210.24, ""text"": ""will be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:07.760"", ""start_sec"": 1207.76, ""end_sec"": 1212.4, ""text"": ""0 up to a certain value and beyond that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:10.240"", ""start_sec"": 1210.24, ""end_sec"": 1214.96, ""text"": ""this is also known as a step function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:12.400"", ""start_sec"": 1212.4, ""end_sec"": 1217.6, ""text"": ""and beyond that it will be 1. in case of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:14.960"", ""start_sec"": 1214.96, ""end_sec"": 1219.52, ""text"": ""sigmoid there is a gradual increase but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:17.600"", ""start_sec"": 1217.6, ""end_sec"": 1222.0, ""text"": ""in case of threshold it's like also"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:19.520"", ""start_sec"": 1219.52, ""end_sec"": 1224.4, ""text"": ""known as a step function there's a rapid"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:22.000"", ""start_sec"": 1222.0, ""end_sec"": 1226.08, ""text"": ""or instantaneous change from zero to one"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:24.400"", ""start_sec"": 1224.4, ""end_sec"": 1228.4, ""text"": ""whereas in sigmoid we will see in the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:26.080"", ""start_sec"": 1226.08, ""end_sec"": 1230.64, ""text"": ""next slide there is a gradual increase"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:28.400"", ""start_sec"": 1228.4, ""end_sec"": 1233.2, ""text"": ""but the value in this case is between"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:30.640"", ""start_sec"": 1230.64, ""end_sec"": 1235.6, ""text"": ""zero and one as well now relu function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:33.200"", ""start_sec"": 1233.2, ""end_sec"": 1238.88, ""text"": ""on the other hand it is equal to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:35.600"", ""start_sec"": 1235.6, ""end_sec"": 1242.96, ""text"": ""basically if the input is 0 or less than"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:38.880"", ""start_sec"": 1238.88, ""end_sec"": 1246.0, ""text"": ""0 then the output is 0 whereas if the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:42.960"", ""start_sec"": 1242.96, ""end_sec"": 1248.0, ""text"": ""input is greater than 0 then the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:46.000"", ""start_sec"": 1246.0, ""end_sec"": 1249.919, ""text"": ""is equal to the input i know it's a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:48.000"", ""start_sec"": 1248.0, ""end_sec"": 1252.4, ""text"": ""little confusing but in the next slides"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:49.919"", ""start_sec"": 1249.919, ""end_sec"": 1254.72, ""text"": ""where we show the relu function it will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:52.400"", ""start_sec"": 1252.4, ""end_sec"": 1257.679, ""text"": ""become clear similarly hyperbolic"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:54.720"", ""start_sec"": 1254.72, ""end_sec"": 1260.159, ""text"": ""tangent this is similar to sigmoid in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:20:57.679"", ""start_sec"": 1257.679, ""end_sec"": 1263.36, ""text"": ""terms of the shape of the function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:00.159"", ""start_sec"": 1260.159, ""end_sec"": 1266.4, ""text"": ""however while sigmoid goes from 0 to 1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:03.360"", ""start_sec"": 1263.36, ""end_sec"": 1269.52, ""text"": ""hyperbolic tangent goes from -1 to 1 and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:06.400"", ""start_sec"": 1266.4, ""end_sec"": 1273.76, ""text"": ""here again the increase or the change"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:09.520"", ""start_sec"": 1269.52, ""end_sec"": 1275.76, ""text"": ""from -1 to 1 is gradual and not like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:13.760"", ""start_sec"": 1273.76, ""end_sec"": 1278.08, ""text"": ""threshold or step function where it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:15.760"", ""start_sec"": 1275.76, ""end_sec"": 1280.159, ""text"": ""happens instantaneously so let's take a"", ""video_id"": ""FbxTVRfQFuI""}]",1177.6,1280.159
little detailed look at some of thesefunctionsso let's start with the sigmoidfunctionso this is the equation of a sigmoidfunctionwhich is 1 by 1 plus e to the power of minusxso x is thevaluethat is theinputit goes from 0 to-1so this is sigmoidfunctionthe equation isphix is equal to 1 by 1 plus e to the power of minus xandas you can seeherethis is the input on thex-axisas x is where the value is comingfrominfactit can also gonegativethis is negativeactuallyso this is thezeroso this is the negative value ofxsoas x is coming from negative value towardszerothe value of theoutput slowlyas it is approachingzeroit it slowly and very gentlyincreasesandactuallyat thepointlet me just use apenat the point here itisit is0.5it is actually0.5 okayandslowly graduallyit increases to 1 as the value of xincreasesbutthenas the value of xincreasesit tapersoffit doesn't go beyondoneso that is the speciality of sigmoidfunctionso the output value will remain between zero andoneit will never go below zero or aboveoneokaythenso that is sigmoidfunctionnow this is thresholdfunctionor this is also referred to as a stepfunctionand here we can also set thethresholdin this case itisthat's why it's called the thresholdfunctionnormally it is0but you can also set a different value for thethresholdnow the difference between this and the sigmoid is that here the change is rapid orinstantaneousas the x value comes from negative up tozeroit remainszero,FbxTVRfQFuI,00:21:18.080,"[{""timestamp"": ""00:21:18.080"", ""start_sec"": 1278.08, ""end_sec"": 1281.919, ""text"": ""little detailed look at some of these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:20.159"", ""start_sec"": 1280.159, ""end_sec"": 1283.919, ""text"": ""functions so let's start with the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:21.919"", ""start_sec"": 1281.919, ""end_sec"": 1286.559, ""text"": ""sigmoid function so this is the equation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:23.919"", ""start_sec"": 1283.919, ""end_sec"": 1289.679, ""text"": ""of a sigmoid function which is 1 by 1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:26.559"", ""start_sec"": 1286.559, ""end_sec"": 1292.799, ""text"": ""plus e to the power of minus x so x is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:29.679"", ""start_sec"": 1289.679, ""end_sec"": 1296.88, ""text"": ""the value that is the input it goes from"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:32.799"", ""start_sec"": 1292.799, ""end_sec"": 1300.0, ""text"": ""0 to -1 so this is sigmoid function the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:36.880"", ""start_sec"": 1296.88, ""end_sec"": 1302.64, ""text"": ""equation is phi x is equal to 1 by 1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:40.000"", ""start_sec"": 1300.0, ""end_sec"": 1304.4, ""text"": ""plus e to the power of minus x and as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:42.640"", ""start_sec"": 1302.64, ""end_sec"": 1307.52, ""text"": ""you can see here this is the input on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:44.400"", ""start_sec"": 1304.4, ""end_sec"": 1309.6, ""text"": ""the x-axis as x is where the value is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:47.520"", ""start_sec"": 1307.52, ""end_sec"": 1311.44, ""text"": ""coming from in fact it can also go"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:49.600"", ""start_sec"": 1309.6, ""end_sec"": 1313.2, ""text"": ""negative this is negative actually so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:51.440"", ""start_sec"": 1311.44, ""end_sec"": 1315.52, ""text"": ""this is the zero so this is the negative"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:53.200"", ""start_sec"": 1313.2, ""end_sec"": 1318.72, ""text"": ""value of x so as x is coming from"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:55.520"", ""start_sec"": 1315.52, ""end_sec"": 1322.08, ""text"": ""negative value towards zero the value of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:21:58.720"", ""start_sec"": 1318.72, ""end_sec"": 1325.12, ""text"": ""the output slowly as it is approaching"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:02.080"", ""start_sec"": 1322.08, ""end_sec"": 1328.32, ""text"": ""zero it it slowly and very gently"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:05.120"", ""start_sec"": 1325.12, ""end_sec"": 1331.6, ""text"": ""increases and actually at the point let"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:08.320"", ""start_sec"": 1328.32, ""end_sec"": 1335.919, ""text"": ""me just use a pen at the point here it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:11.600"", ""start_sec"": 1331.6, ""end_sec"": 1339.039, ""text"": ""is it is 0.5 it is actually 0.5 okay and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:15.919"", ""start_sec"": 1335.919, ""end_sec"": 1341.44, ""text"": ""slowly gradually it increases to 1 as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:19.039"", ""start_sec"": 1339.039, ""end_sec"": 1344.4, ""text"": ""the value of x increases but then as the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:21.440"", ""start_sec"": 1341.44, ""end_sec"": 1347.36, ""text"": ""value of x increases it tapers off it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:24.400"", ""start_sec"": 1344.4, ""end_sec"": 1349.84, ""text"": ""doesn't go beyond one so that is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:27.360"", ""start_sec"": 1347.36, ""end_sec"": 1352.32, ""text"": ""speciality of sigmoid function so the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:29.840"", ""start_sec"": 1349.84, ""end_sec"": 1354.96, ""text"": ""output value will remain between zero"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:32.320"", ""start_sec"": 1352.32, ""end_sec"": 1357.36, ""text"": ""and one it will never go below zero or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:34.960"", ""start_sec"": 1354.96, ""end_sec"": 1359.679, ""text"": ""above one okay then so that is sigmoid"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:37.360"", ""start_sec"": 1357.36, ""end_sec"": 1362.0, ""text"": ""function now this is threshold function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:39.679"", ""start_sec"": 1359.679, ""end_sec"": 1364.88, ""text"": ""or this is also referred to as a step"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:42.000"", ""start_sec"": 1362.0, ""end_sec"": 1366.64, ""text"": ""function and here we can also set the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:44.880"", ""start_sec"": 1364.88, ""end_sec"": 1368.24, ""text"": ""threshold in this case it is that's why"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:46.640"", ""start_sec"": 1366.64, ""end_sec"": 1370.72, ""text"": ""it's called the threshold function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:48.240"", ""start_sec"": 1368.24, ""end_sec"": 1372.559, ""text"": ""normally it is 0 but you can also set a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:50.720"", ""start_sec"": 1370.72, ""end_sec"": 1374.24, ""text"": ""different value for the threshold now"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:52.559"", ""start_sec"": 1372.559, ""end_sec"": 1377.12, ""text"": ""the difference between this and the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:54.240"", ""start_sec"": 1374.24, ""end_sec"": 1379.84, ""text"": ""sigmoid is that here the change is rapid"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:57.120"", ""start_sec"": 1377.12, ""end_sec"": 1382.799, ""text"": ""or instantaneous as the x value comes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:22:59.840"", ""start_sec"": 1379.84, ""end_sec"": 1386.24, ""text"": ""from negative up to zero it remains zero"", ""video_id"": ""FbxTVRfQFuI""}]",1278.08,1386.24
and at zero it pretty much immediately increases to1okayso this is a mathematical representation of thresholdfunctionphi x is equal to 1 if x is greater than equal to 0 and 0 if x is less than 0. so for all negative values it is0 whichsince we have set the threshold to be0so as soon as it reaches 0 it becomes1.you see the difference between this and the previousonewhich is basically thesigmoidwhere the increase from 0 to 1 is gradual and here it isinstantaneousand that's why this is also known as a stepfunctionthreshold function or stepfunctionthis is arelua relu is one of the most popular activation functionstodaythis is the definition ofreluphi x is equal to max of x commazerowhat it saysisif the value of x is less thanzerothen phi x iszerothe moment itincreasesgoes beyondzerothe value of phi x is equal toxso it doesn't stop atoneactually it goes all thewayso as the value of xincreasesthe value of y will also increaseinfinitelyso there is no limithereunlike your sigmoid or threshold or the nextonewhich is basicallyhyperbolictangentokayso in case ofrelu rememberthere is no upperlimitthe output is equal to either0in case the value of x isnegativeor it is equal to the value ofxsoforexample hereif the value of x is10then the value of y is also10right okayso that isreluand there are several advantages ofreluand it is much more efficient and provides much more accuracy compared to other activation functions like,FbxTVRfQFuI,00:23:02.799,"[{""timestamp"": ""00:23:02.799"", ""start_sec"": 1382.799, ""end_sec"": 1388.64, ""text"": ""and at zero it pretty much immediately"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:06.240"", ""start_sec"": 1386.24, ""end_sec"": 1391.28, ""text"": ""increases to 1 okay so this is a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:08.640"", ""start_sec"": 1388.64, ""end_sec"": 1393.919, ""text"": ""mathematical representation of threshold"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:11.280"", ""start_sec"": 1391.28, ""end_sec"": 1396.799, ""text"": ""function phi x is equal to 1 if x is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:13.919"", ""start_sec"": 1393.919, ""end_sec"": 1398.799, ""text"": ""greater than equal to 0 and 0 if x is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:16.799"", ""start_sec"": 1396.799, ""end_sec"": 1400.64, ""text"": ""less than 0. so for all negative values"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:18.799"", ""start_sec"": 1398.799, ""end_sec"": 1403.12, ""text"": ""it is 0 which since we have set the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:20.640"", ""start_sec"": 1400.64, ""end_sec"": 1405.679, ""text"": ""threshold to be 0 so as soon as it"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:23.120"", ""start_sec"": 1403.12, ""end_sec"": 1408.64, ""text"": ""reaches 0 it becomes 1. you see the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:25.679"", ""start_sec"": 1405.679, ""end_sec"": 1411.52, ""text"": ""difference between this and the previous"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:28.640"", ""start_sec"": 1408.64, ""end_sec"": 1414.72, ""text"": ""one which is basically the sigmoid where"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:31.520"", ""start_sec"": 1411.52, ""end_sec"": 1417.12, ""text"": ""the increase from 0 to 1 is gradual and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:34.720"", ""start_sec"": 1414.72, ""end_sec"": 1419.2, ""text"": ""here it is instantaneous and that's why"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:37.120"", ""start_sec"": 1417.12, ""end_sec"": 1421.44, ""text"": ""this is also known as a step function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:39.200"", ""start_sec"": 1419.2, ""end_sec"": 1423.679, ""text"": ""threshold function or step function this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:41.440"", ""start_sec"": 1421.44, ""end_sec"": 1426.159, ""text"": ""is a relu a relu is one of the most"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:43.679"", ""start_sec"": 1423.679, ""end_sec"": 1428.799, ""text"": ""popular activation functions today this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:46.159"", ""start_sec"": 1426.159, ""end_sec"": 1431.679, ""text"": ""is the definition of relu phi x is equal"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:48.799"", ""start_sec"": 1428.799, ""end_sec"": 1434.4, ""text"": ""to max of x comma zero what it says is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:51.679"", ""start_sec"": 1431.679, ""end_sec"": 1435.679, ""text"": ""if the value of x is less than zero then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:54.400"", ""start_sec"": 1434.4, ""end_sec"": 1438.88, ""text"": ""phi x is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:55.679"", ""start_sec"": 1435.679, ""end_sec"": 1443.6, ""text"": ""zero the moment it increases goes beyond"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:23:58.880"", ""start_sec"": 1438.88, ""end_sec"": 1446.72, ""text"": ""zero the value of phi x is equal to x so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:03.600"", ""start_sec"": 1443.6, ""end_sec"": 1448.799, ""text"": ""it doesn't stop at one actually it goes"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:06.720"", ""start_sec"": 1446.72, ""end_sec"": 1450.72, ""text"": ""all the way so as the value of x"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:08.799"", ""start_sec"": 1448.799, ""end_sec"": 1453.44, ""text"": ""increases the value of y will also"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:10.720"", ""start_sec"": 1450.72, ""end_sec"": 1457.76, ""text"": ""increase infinitely so there is no limit"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:13.440"", ""start_sec"": 1453.44, ""end_sec"": 1459.76, ""text"": ""here unlike your sigmoid or threshold or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:17.760"", ""start_sec"": 1457.76, ""end_sec"": 1462.559, ""text"": ""the next one which is basically"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:19.760"", ""start_sec"": 1459.76, ""end_sec"": 1465.2, ""text"": ""hyperbolic tangent okay so in case of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:22.559"", ""start_sec"": 1462.559, ""end_sec"": 1468.08, ""text"": ""relu remember there is no upper limit"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:25.200"", ""start_sec"": 1465.2, ""end_sec"": 1471.039, ""text"": ""the output is equal to either 0 in case"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:28.080"", ""start_sec"": 1468.08, ""end_sec"": 1474.24, ""text"": ""the value of x is negative or it is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:31.039"", ""start_sec"": 1471.039, ""end_sec"": 1477.039, ""text"": ""equal to the value of x so for example"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:34.240"", ""start_sec"": 1474.24, ""end_sec"": 1479.84, ""text"": ""here if the value of x is 10 then the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:37.039"", ""start_sec"": 1477.039, ""end_sec"": 1482.96, ""text"": ""value of y is also 10 right okay so that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:39.840"", ""start_sec"": 1479.84, ""end_sec"": 1485.679, ""text"": ""is relu and there are several advantages"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:42.960"", ""start_sec"": 1482.96, ""end_sec"": 1488.159, ""text"": ""of relu and it is much more efficient"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:45.679"", ""start_sec"": 1485.679, ""end_sec"": 1489.84, ""text"": ""and provides much more accuracy compared"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:48.159"", ""start_sec"": 1488.159, ""end_sec"": 1491.679, ""text"": ""to other activation functions like"", ""video_id"": ""FbxTVRfQFuI""}]",1382.799,1491.679
sigmoid and soonso that's the reason it is verypopularallrightso this is hyperbolic tangent activationfunctionthe function looks similar to sigmoidfunctionthecurveif you see theshapeit looks similar to sigmoidfunctionbut the difference between hyperbolic tangent and sigmoid function is that in case of sigmoid the output goes from zero toonewhereas in case of hyperbolic tangent it goes from -1 to1so that is the difference between hyperbolic tangent and sigmoidfunctionotherwisethe shape looks verysimilarthere is a gradualincreaseunlike the step function where there was an instant increase or instantchangehereagain very similar to sigmoidfunctionthe value changes gradually from -1 to1.so this is the equation of hyperbolic tangent activationfunctionyeahso then let's moveonthis is a diagrammatic representation of the activation function and how the overalldatahow the overallprogressionhappens from input to theoutputso we get the input from the inputlayerby thewaythe neural network has threelayerstypically there will be threelayersthere is an inputlayerthere is an output layer and then you have the hiddenlayerso the inputs come from the input layer and they get processed in the hidden layer and then you get the output in the outputlayerso let's take a little bit of a detailed look into the working of a neuralnetworkso let's say we want to classify some images between dogs andcatshow do we dothisthis is known as a classification process and we,FbxTVRfQFuI,00:24:49.840,"[{""timestamp"": ""00:24:49.840"", ""start_sec"": 1489.84, ""end_sec"": 1494.32, ""text"": ""sigmoid and so on so that's the reason"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:51.679"", ""start_sec"": 1491.679, ""end_sec"": 1496.64, ""text"": ""it is very popular all right so this is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:54.320"", ""start_sec"": 1494.32, ""end_sec"": 1498.64, ""text"": ""hyperbolic tangent activation function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:56.640"", ""start_sec"": 1496.64, ""end_sec"": 1501.279, ""text"": ""the function looks similar to sigmoid"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:24:58.640"", ""start_sec"": 1498.64, ""end_sec"": 1503.36, ""text"": ""function the curve if you see the shape"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:01.279"", ""start_sec"": 1501.279, ""end_sec"": 1505.279, ""text"": ""it looks similar to sigmoid function but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:03.360"", ""start_sec"": 1503.36, ""end_sec"": 1508.08, ""text"": ""the difference between hyperbolic"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:05.279"", ""start_sec"": 1505.279, ""end_sec"": 1510.799, ""text"": ""tangent and sigmoid function is that in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:08.080"", ""start_sec"": 1508.08, ""end_sec"": 1513.2, ""text"": ""case of sigmoid the output goes from"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:10.799"", ""start_sec"": 1510.799, ""end_sec"": 1516.96, ""text"": ""zero to one whereas in case of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:13.200"", ""start_sec"": 1513.2, ""end_sec"": 1518.559, ""text"": ""hyperbolic tangent it goes from -1 to 1"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:16.960"", ""start_sec"": 1516.96, ""end_sec"": 1521.36, ""text"": ""so that is the difference between"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:18.559"", ""start_sec"": 1518.559, ""end_sec"": 1523.84, ""text"": ""hyperbolic tangent and sigmoid function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:21.360"", ""start_sec"": 1521.36, ""end_sec"": 1526.799, ""text"": ""otherwise the shape looks very similar"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:23.840"", ""start_sec"": 1523.84, ""end_sec"": 1529.279, ""text"": ""there is a gradual increase unlike the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:26.799"", ""start_sec"": 1526.799, ""end_sec"": 1531.84, ""text"": ""step function where there was an instant"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:29.279"", ""start_sec"": 1529.279, ""end_sec"": 1534.159, ""text"": ""increase or instant change here again"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:31.840"", ""start_sec"": 1531.84, ""end_sec"": 1537.679, ""text"": ""very similar to sigmoid function the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:34.159"", ""start_sec"": 1534.159, ""end_sec"": 1540.08, ""text"": ""value changes gradually from -1 to 1. so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:37.679"", ""start_sec"": 1537.679, ""end_sec"": 1542.72, ""text"": ""this is the equation of hyperbolic"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:40.080"", ""start_sec"": 1540.08, ""end_sec"": 1544.799, ""text"": ""tangent activation function yeah so then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:42.720"", ""start_sec"": 1542.72, ""end_sec"": 1547.2, ""text"": ""let's move on this is a diagrammatic"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:44.799"", ""start_sec"": 1544.799, ""end_sec"": 1550.88, ""text"": ""representation of the activation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:47.200"", ""start_sec"": 1547.2, ""end_sec"": 1553.44, ""text"": ""function and how the overall data how"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:50.880"", ""start_sec"": 1550.88, ""end_sec"": 1555.84, ""text"": ""the overall progression happens from"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:53.440"", ""start_sec"": 1553.44, ""end_sec"": 1557.679, ""text"": ""input to the output so we get the input"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:55.840"", ""start_sec"": 1555.84, ""end_sec"": 1559.919, ""text"": ""from the input layer by the way the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:57.679"", ""start_sec"": 1557.679, ""end_sec"": 1561.44, ""text"": ""neural network has three layers"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:25:59.919"", ""start_sec"": 1559.919, ""end_sec"": 1563.12, ""text"": ""typically there will be three layers"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:01.440"", ""start_sec"": 1561.44, ""end_sec"": 1564.88, ""text"": ""there is an input layer there is an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:03.120"", ""start_sec"": 1563.12, ""end_sec"": 1567.6, ""text"": ""output layer and then you have the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:04.880"", ""start_sec"": 1564.88, ""end_sec"": 1570.24, ""text"": ""hidden layer so the inputs come from the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:07.600"", ""start_sec"": 1567.6, ""end_sec"": 1572.24, ""text"": ""input layer and they get processed in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:10.240"", ""start_sec"": 1570.24, ""end_sec"": 1574.4, ""text"": ""the hidden layer and then you get the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:12.240"", ""start_sec"": 1572.24, ""end_sec"": 1576.96, ""text"": ""output in the output layer so let's take"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:14.400"", ""start_sec"": 1574.4, ""end_sec"": 1579.84, ""text"": ""a little bit of a detailed look into the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:16.960"", ""start_sec"": 1576.96, ""end_sec"": 1582.88, ""text"": ""working of a neural network so let's say"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:19.840"", ""start_sec"": 1579.84, ""end_sec"": 1585.679, ""text"": ""we want to classify some images between"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:22.880"", ""start_sec"": 1582.88, ""end_sec"": 1588.4, ""text"": ""dogs and cats how do we do this this is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:25.679"", ""start_sec"": 1585.679, ""end_sec"": 1590.159, ""text"": ""known as a classification process and we"", ""video_id"": ""FbxTVRfQFuI""}]",1489.84,1590.159
are trying to use neural networks and deep learning to implement thisclassificationso how do we dothatso this is how itworksso you have four layer neuralnetworkthere is an inputlayerthere is an output layer and then there are two hiddenlayersand what we do is we provide labeled trainingdatawhich means these images are fed to the network with the label sayingthat okaythis is acatthe neural network is allowed to process it and come up with a prediction saying whether it is a cat or adogand obviously in the beginning there may bemistakesa cat may be classified as adogso we then saythat okaythis iswrongthis output iswrongbut every time it predictscorrectlywesay yesthis output iscorrectso that learningprocessso it will gobackmake some changes to its weights andbiaseswe again feed these inputs and it will give us theoutputwe will check whether it is correct ornotand soonso this is a iterativeprocesswhich is known as the trainingprocessso we are training the neuralnetworkand what happens in the trainingprocessthese weights andbiasesyou remember there were weights likew1w2 and soonso these weights and biases keep changing every time you feedthesewhich is known as anepochso there are multipleiterationsevery iteration is known as an epoch and each time the weights are dated to make sure that the maximum number of images are classifiedcorrectlysoonceagainwhat is theinputthis input could be like 1000 images of cats and dogs and they are labeled because we know which is a,FbxTVRfQFuI,00:26:28.400,"[{""timestamp"": ""00:26:28.400"", ""start_sec"": 1588.4, ""end_sec"": 1591.6, ""text"": ""are trying to use neural networks and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:30.159"", ""start_sec"": 1590.159, ""end_sec"": 1593.52, ""text"": ""deep learning to implement this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:31.600"", ""start_sec"": 1591.6, ""end_sec"": 1597.44, ""text"": ""classification so how do we do that so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:33.520"", ""start_sec"": 1593.52, ""end_sec"": 1600.159, ""text"": ""this is how it works so you have four"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:37.440"", ""start_sec"": 1597.44, ""end_sec"": 1602.559, ""text"": ""layer neural network there is an input"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:40.159"", ""start_sec"": 1600.159, ""end_sec"": 1605.44, ""text"": ""layer there is an output layer and then"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:42.559"", ""start_sec"": 1602.559, ""end_sec"": 1609.44, ""text"": ""there are two hidden layers and what we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:45.440"", ""start_sec"": 1605.44, ""end_sec"": 1612.08, ""text"": ""do is we provide labeled training data"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:49.440"", ""start_sec"": 1609.44, ""end_sec"": 1614.64, ""text"": ""which means these images are fed to the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:52.080"", ""start_sec"": 1612.08, ""end_sec"": 1617.12, ""text"": ""network with the label saying that okay"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:54.640"", ""start_sec"": 1614.64, ""end_sec"": 1620.159, ""text"": ""this is a cat the neural network is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:26:57.120"", ""start_sec"": 1617.12, ""end_sec"": 1622.48, ""text"": ""allowed to process it and come up with a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:00.159"", ""start_sec"": 1620.159, ""end_sec"": 1625.039, ""text"": ""prediction saying whether it is a cat or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:02.480"", ""start_sec"": 1622.48, ""end_sec"": 1627.2, ""text"": ""a dog and obviously in the beginning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:05.039"", ""start_sec"": 1625.039, ""end_sec"": 1629.76, ""text"": ""there may be mistakes a cat may be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:07.200"", ""start_sec"": 1627.2, ""end_sec"": 1632.08, ""text"": ""classified as a dog so we then say that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:09.760"", ""start_sec"": 1629.76, ""end_sec"": 1634.0, ""text"": ""okay this is wrong this output is wrong"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:12.080"", ""start_sec"": 1632.08, ""end_sec"": 1636.559, ""text"": ""but every time it predicts correctly we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:14.000"", ""start_sec"": 1634.0, ""end_sec"": 1639.12, ""text"": ""say yes this output is correct so that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:16.559"", ""start_sec"": 1636.559, ""end_sec"": 1641.76, ""text"": ""learning process so it will go back make"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:19.120"", ""start_sec"": 1639.12, ""end_sec"": 1644.72, ""text"": ""some changes to its weights and biases"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:21.760"", ""start_sec"": 1641.76, ""end_sec"": 1646.799, ""text"": ""we again feed these inputs and it will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:24.720"", ""start_sec"": 1644.72, ""end_sec"": 1648.799, ""text"": ""give us the output we will check whether"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:26.799"", ""start_sec"": 1646.799, ""end_sec"": 1651.36, ""text"": ""it is correct or not and so on so this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:28.799"", ""start_sec"": 1648.799, ""end_sec"": 1654.32, ""text"": ""is a iterative process which is known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:31.360"", ""start_sec"": 1651.36, ""end_sec"": 1656.88, ""text"": ""the training process so we are training"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:34.320"", ""start_sec"": 1654.32, ""end_sec"": 1659.44, ""text"": ""the neural network and what happens in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:36.880"", ""start_sec"": 1656.88, ""end_sec"": 1661.76, ""text"": ""the training process these weights and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:39.440"", ""start_sec"": 1659.44, ""end_sec"": 1665.6, ""text"": ""biases you remember there were weights"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:41.760"", ""start_sec"": 1661.76, ""end_sec"": 1668.88, ""text"": ""like w1 w2 and so on so these weights"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:45.600"", ""start_sec"": 1665.6, ""end_sec"": 1671.679, ""text"": ""and biases keep changing every time you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:48.880"", ""start_sec"": 1668.88, ""end_sec"": 1673.76, ""text"": ""feed these which is known as an epoch so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:51.679"", ""start_sec"": 1671.679, ""end_sec"": 1676.159, ""text"": ""there are multiple iterations every"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:53.760"", ""start_sec"": 1673.76, ""end_sec"": 1678.96, ""text"": ""iteration is known as an epoch and each"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:56.159"", ""start_sec"": 1676.159, ""end_sec"": 1681.279, ""text"": ""time the weights are dated to make sure"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:27:58.960"", ""start_sec"": 1678.96, ""end_sec"": 1683.679, ""text"": ""that the maximum number of images are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:01.279"", ""start_sec"": 1681.279, ""end_sec"": 1686.08, ""text"": ""classified correctly so once again what"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:03.679"", ""start_sec"": 1683.679, ""end_sec"": 1689.6, ""text"": ""is the input this input could be like"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:06.080"", ""start_sec"": 1686.08, ""end_sec"": 1692.159, ""text"": ""1000 images of cats and dogs and they"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:09.600"", ""start_sec"": 1689.6, ""end_sec"": 1694.559, ""text"": ""are labeled because we know which is a"", ""video_id"": ""FbxTVRfQFuI""}]",1588.4,1694.559
cat and which is a dog and we feed those thousandimagesthe neural network will initially assign some weights and biases for each neuron and it will try toprocessextract the features from the images and it will try to come up with a prediction for eachimageand thatpredictionthat is calculated by thenetworkis compared with the actualvaluewhether it is a cat or adogand that's how the error iscalculatedso let's say there are a thousand images and in the first run only 500 of them have been correctlyclassifiedthat means we are getting only 50accuracyso we feed that information back to thenetworkfurther update these weights and biases for each of the neurons and we runthistheseinputsonceagainit will try tocalculateextract the features and it will try to predict which of these is cats and dogsand thistimelet'ssayout ofthousand700 of them have been predictedcorrectlyso that means in the second iteration the accuracy has increased from 50 to 70percentallrightthen we go backagainwe feedthismaybe for a thirditerationfourth iteration and soonand slowly and steadily the accuracy of this network will keep increasing and it may reachprobablyyou neverknow9095percentand there are several parameters that are known as hyper parameters that need to be changed andtweakedand that is the overall trainingprocessand ultimately at some point wesay okayyou will probably never reach hundred percentaccuracybut then we set alimitsaying thatokayif we receive95,FbxTVRfQFuI,00:28:12.159,"[{""timestamp"": ""00:28:12.159"", ""start_sec"": 1692.159, ""end_sec"": 1697.039, ""text"": ""cat and which is a dog and we feed those"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:14.559"", ""start_sec"": 1694.559, ""end_sec"": 1698.96, ""text"": ""thousand images the neural network will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:17.039"", ""start_sec"": 1697.039, ""end_sec"": 1700.799, ""text"": ""initially assign some weights and biases"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:18.960"", ""start_sec"": 1698.96, ""end_sec"": 1703.12, ""text"": ""for each neuron and it will try to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:20.799"", ""start_sec"": 1700.799, ""end_sec"": 1705.12, ""text"": ""process extract the features from the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:23.120"", ""start_sec"": 1703.12, ""end_sec"": 1707.279, ""text"": ""images and it will try to come up with a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:25.120"", ""start_sec"": 1705.12, ""end_sec"": 1709.679, ""text"": ""prediction for each image and that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:27.279"", ""start_sec"": 1707.279, ""end_sec"": 1712.24, ""text"": ""prediction that is calculated by the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:29.679"", ""start_sec"": 1709.679, ""end_sec"": 1714.24, ""text"": ""network is compared with the actual"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:32.240"", ""start_sec"": 1712.24, ""end_sec"": 1716.399, ""text"": ""value whether it is a cat or a dog and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:34.240"", ""start_sec"": 1714.24, ""end_sec"": 1718.559, ""text"": ""that's how the error is calculated so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:36.399"", ""start_sec"": 1716.399, ""end_sec"": 1721.279, ""text"": ""let's say there are a thousand images"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:38.559"", ""start_sec"": 1718.559, ""end_sec"": 1723.2, ""text"": ""and in the first run only 500 of them"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:41.279"", ""start_sec"": 1721.279, ""end_sec"": 1725.44, ""text"": ""have been correctly classified that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:43.200"", ""start_sec"": 1723.2, ""end_sec"": 1727.44, ""text"": ""means we are getting only 50 accuracy so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:45.440"", ""start_sec"": 1725.44, ""end_sec"": 1729.76, ""text"": ""we feed that information back to the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:47.440"", ""start_sec"": 1727.44, ""end_sec"": 1731.919, ""text"": ""network further update these weights and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:49.760"", ""start_sec"": 1729.76, ""end_sec"": 1734.48, ""text"": ""biases for each of the neurons and we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:51.919"", ""start_sec"": 1731.919, ""end_sec"": 1736.32, ""text"": ""run this these inputs once again it will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:54.480"", ""start_sec"": 1734.48, ""end_sec"": 1738.0, ""text"": ""try to calculate extract the features"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:56.320"", ""start_sec"": 1736.32, ""end_sec"": 1739.84, ""text"": ""and it will try to predict which of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:58.000"", ""start_sec"": 1738.0, ""end_sec"": 1742.399, ""text"": ""these is cats and dogs and this time"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:28:59.840"", ""start_sec"": 1739.84, ""end_sec"": 1744.48, ""text"": ""let's say out of thousand 700 of them"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:02.399"", ""start_sec"": 1742.399, ""end_sec"": 1746.72, ""text"": ""have been predicted correctly so that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:04.480"", ""start_sec"": 1744.48, ""end_sec"": 1749.679, ""text"": ""means in the second iteration the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:06.720"", ""start_sec"": 1746.72, ""end_sec"": 1752.559, ""text"": ""accuracy has increased from 50 to 70"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:09.679"", ""start_sec"": 1749.679, ""end_sec"": 1755.039, ""text"": ""percent all right then we go back again"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:12.559"", ""start_sec"": 1752.559, ""end_sec"": 1757.76, ""text"": ""we feed this maybe for a third iteration"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:15.039"", ""start_sec"": 1755.039, ""end_sec"": 1760.799, ""text"": ""fourth iteration and so on and slowly"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:17.760"", ""start_sec"": 1757.76, ""end_sec"": 1763.36, ""text"": ""and steadily the accuracy of this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:20.799"", ""start_sec"": 1760.799, ""end_sec"": 1766.08, ""text"": ""network will keep increasing and it may"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:23.360"", ""start_sec"": 1763.36, ""end_sec"": 1768.24, ""text"": ""reach probably you never know 90 95"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:26.080"", ""start_sec"": 1766.08, ""end_sec"": 1770.24, ""text"": ""percent and there are several parameters"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:28.240"", ""start_sec"": 1768.24, ""end_sec"": 1772.72, ""text"": ""that are known as hyper parameters that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:30.240"", ""start_sec"": 1770.24, ""end_sec"": 1774.88, ""text"": ""need to be changed and tweaked and that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:32.720"", ""start_sec"": 1772.72, ""end_sec"": 1777.76, ""text"": ""is the overall training process and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:34.880"", ""start_sec"": 1774.88, ""end_sec"": 1779.2, ""text"": ""ultimately at some point we say okay you"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:37.760"", ""start_sec"": 1777.76, ""end_sec"": 1782.08, ""text"": ""will probably never reach hundred"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:39.200"", ""start_sec"": 1779.2, ""end_sec"": 1784.159, ""text"": ""percent accuracy but then we set a limit"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:42.080"", ""start_sec"": 1782.08, ""end_sec"": 1786.08, ""text"": ""saying that okay if we receive 95"", ""video_id"": ""FbxTVRfQFuI""}]",1692.159,1786.08
percentaccuracythat is good enough for ourapplicationand then wesay okayour training process isdoneso that is the way traininghappensand once the training isdonenow with the training datasetthe systemhaslet'ssayseen all these thousandimagesthereforewhat we do is the nextsteplike in any normal machine learningprocesswe do thetestingwhere we take a fresh set of images and we feed it to thenetworkthe fresh set which it has not seenbeforeas a part of the trainingprocessand this isagainnothing new in deeplearningthis was there in machine learning aswellso you feed the test images and then find out whether we are getting a similar accuracy ornotso maybe that accuracy may reduce a littlebitwhile training you may get 98 percent and then for test you may get 95percentbut there shouldn't be a drasticdroplikeforexampleyou get 98 in training and then you get 50 or 40 percent with thetestthat means your network has notlearnedyou may have to retrain yournetworkso that is the way neural network trainingworksandrememberthe whole process is about changing these weights and biases and coming up with the optimal values of these weights and biases so that the accuracy is the maximumpossibleallrightso a little bit more detail about how this whole thingworksso this is known as forwardpropagationwhich is the data or the information is going in the forwarddirectionthe inputs aretaken weightedsummation isdonebias is added here and then that is fed to the,FbxTVRfQFuI,00:29:44.159,"[{""timestamp"": ""00:29:44.159"", ""start_sec"": 1784.159, ""end_sec"": 1788.399, ""text"": ""percent accuracy that is good enough for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:46.080"", ""start_sec"": 1786.08, ""end_sec"": 1790.32, ""text"": ""our application and then we say okay our"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:48.399"", ""start_sec"": 1788.399, ""end_sec"": 1793.12, ""text"": ""training process is done so that is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:50.320"", ""start_sec"": 1790.32, ""end_sec"": 1795.76, ""text"": ""way training happens and once the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:53.120"", ""start_sec"": 1793.12, ""end_sec"": 1798.399, ""text"": ""training is done now with the training"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:55.760"", ""start_sec"": 1795.76, ""end_sec"": 1801.039, ""text"": ""data set the system has let's say seen"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:29:58.399"", ""start_sec"": 1798.399, ""end_sec"": 1803.76, ""text"": ""all these thousand images therefore what"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:01.039"", ""start_sec"": 1801.039, ""end_sec"": 1805.84, ""text"": ""we do is the next step like in any"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:03.760"", ""start_sec"": 1803.76, ""end_sec"": 1808.399, ""text"": ""normal machine learning process we do"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:05.840"", ""start_sec"": 1805.84, ""end_sec"": 1810.799, ""text"": ""the testing where we take a fresh set of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:08.399"", ""start_sec"": 1808.399, ""end_sec"": 1813.039, ""text"": ""images and we feed it to the network the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:10.799"", ""start_sec"": 1810.799, ""end_sec"": 1814.88, ""text"": ""fresh set which it has not seen before"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:13.039"", ""start_sec"": 1813.039, ""end_sec"": 1816.559, ""text"": ""as a part of the training process and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:14.880"", ""start_sec"": 1814.88, ""end_sec"": 1818.159, ""text"": ""this is again nothing new in deep"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:16.559"", ""start_sec"": 1816.559, ""end_sec"": 1820.72, ""text"": ""learning this was there in machine"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:18.159"", ""start_sec"": 1818.159, ""end_sec"": 1823.44, ""text"": ""learning as well so you feed the test"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:20.720"", ""start_sec"": 1820.72, ""end_sec"": 1825.52, ""text"": ""images and then find out whether we are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:23.440"", ""start_sec"": 1823.44, ""end_sec"": 1827.6, ""text"": ""getting a similar accuracy or not so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:25.520"", ""start_sec"": 1825.52, ""end_sec"": 1829.52, ""text"": ""maybe that accuracy may reduce a little"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:27.600"", ""start_sec"": 1827.6, ""end_sec"": 1831.84, ""text"": ""bit while training you may get 98"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:29.520"", ""start_sec"": 1829.52, ""end_sec"": 1833.76, ""text"": ""percent and then for test you may get 95"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:31.840"", ""start_sec"": 1831.84, ""end_sec"": 1836.48, ""text"": ""percent but there shouldn't be a drastic"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:33.760"", ""start_sec"": 1833.76, ""end_sec"": 1838.88, ""text"": ""drop like for example you get 98 in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:36.480"", ""start_sec"": 1836.48, ""end_sec"": 1840.799, ""text"": ""training and then you get 50 or 40"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:38.880"", ""start_sec"": 1838.88, ""end_sec"": 1843.279, ""text"": ""percent with the test that means your"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:40.799"", ""start_sec"": 1840.799, ""end_sec"": 1846.32, ""text"": ""network has not learned you may have to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:43.279"", ""start_sec"": 1843.279, ""end_sec"": 1847.919, ""text"": ""retrain your network so that is the way"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:46.320"", ""start_sec"": 1846.32, ""end_sec"": 1850.799, ""text"": ""neural network training works and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:47.919"", ""start_sec"": 1847.919, ""end_sec"": 1853.279, ""text"": ""remember the whole process is about"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:50.799"", ""start_sec"": 1850.799, ""end_sec"": 1855.679, ""text"": ""changing these weights and biases and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:53.279"", ""start_sec"": 1853.279, ""end_sec"": 1857.52, ""text"": ""coming up with the optimal values of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:55.679"", ""start_sec"": 1855.679, ""end_sec"": 1860.24, ""text"": ""these weights and biases so that the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:30:57.520"", ""start_sec"": 1857.52, ""end_sec"": 1862.96, ""text"": ""accuracy is the maximum possible all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:00.240"", ""start_sec"": 1860.24, ""end_sec"": 1864.96, ""text"": ""right so a little bit more detail about"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:02.960"", ""start_sec"": 1862.96, ""end_sec"": 1867.52, ""text"": ""how this whole thing works so this is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:04.960"", ""start_sec"": 1864.96, ""end_sec"": 1869.84, ""text"": ""known as forward propagation which is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:07.520"", ""start_sec"": 1867.52, ""end_sec"": 1872.32, ""text"": ""the data or the information is going in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:09.840"", ""start_sec"": 1869.84, ""end_sec"": 1875.279, ""text"": ""the forward direction the inputs are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:12.320"", ""start_sec"": 1872.32, ""end_sec"": 1878.399, ""text"": ""taken weighted summation is done bias is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:15.279"", ""start_sec"": 1875.279, ""end_sec"": 1881.039, ""text"": ""added here and then that is fed to the"", ""video_id"": ""FbxTVRfQFuI""}]",1784.159,1881.039
activation function and then that is that comes out as anoutputso that is forwardpropagationand the output is compared with the actual value and that will give us theerrorthe difference between them is theerrorand in technical terms that is also known as our costfunctionand this is what we would like tominimizethere are different ways of defining the costfunctionbut one of the simplest ways is mean squareerrorso it is nothing but the square of the difference of theerrorsor the sum of the squares of the difference of theerrorsand this is also nothing new wehaveprobablyif you're familiar with machinelearningyou must have come across this meansquarenow there are different ways of defining costfunctionit need not always be the mean squareerrorbut the most common one isthisso you define this cost function and you ask the system to minimize thiserrorso we use what is known as an optimization function to minimize this error and the error itself sent back to the system asfeedbackand that is known as backpropagationand so this is the costfunctionand how do we optimize the costfunctionwe use what is known as gradientdescentso the gradient descent mechanism identifies how to change the weights and biases so that the cost function isminimizedand there is also what is known as the rate or the learningratethat is what is shown here as slower andfasterso you need to specify what should be the learningratenowif the learning rate is verysmall,FbxTVRfQFuI,00:31:18.399,"[{""timestamp"": ""00:31:18.399"", ""start_sec"": 1878.399, ""end_sec"": 1883.2, ""text"": ""activation function and then that is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:21.039"", ""start_sec"": 1881.039, ""end_sec"": 1885.36, ""text"": ""that comes out as an output so that is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:23.200"", ""start_sec"": 1883.2, ""end_sec"": 1887.36, ""text"": ""forward propagation and the output is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:25.360"", ""start_sec"": 1885.36, ""end_sec"": 1889.039, ""text"": ""compared with the actual value and that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:27.360"", ""start_sec"": 1887.36, ""end_sec"": 1891.2, ""text"": ""will give us the error the difference"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:29.039"", ""start_sec"": 1889.039, ""end_sec"": 1893.679, ""text"": ""between them is the error and in"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:31.200"", ""start_sec"": 1891.2, ""end_sec"": 1896.72, ""text"": ""technical terms that is also known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:33.679"", ""start_sec"": 1893.679, ""end_sec"": 1898.88, ""text"": ""our cost function and this is what we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:36.720"", ""start_sec"": 1896.72, ""end_sec"": 1900.559, ""text"": ""would like to minimize there are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:38.880"", ""start_sec"": 1898.88, ""end_sec"": 1904.0, ""text"": ""different ways of defining the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:40.559"", ""start_sec"": 1900.559, ""end_sec"": 1907.2, ""text"": ""function but one of the simplest ways is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:44.000"", ""start_sec"": 1904.0, ""end_sec"": 1909.12, ""text"": ""mean square error so it is nothing but"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:47.200"", ""start_sec"": 1907.2, ""end_sec"": 1911.919, ""text"": ""the square of the difference of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:49.120"", ""start_sec"": 1909.12, ""end_sec"": 1913.679, ""text"": ""errors or the sum of the squares of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:51.919"", ""start_sec"": 1911.919, ""end_sec"": 1916.24, ""text"": ""difference of the errors and this is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:53.679"", ""start_sec"": 1913.679, ""end_sec"": 1917.76, ""text"": ""also nothing new we have probably if"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:56.240"", ""start_sec"": 1916.24, ""end_sec"": 1919.76, ""text"": ""you're familiar with machine learning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:57.760"", ""start_sec"": 1917.76, ""end_sec"": 1922.159, ""text"": ""you must have come across this mean"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:31:59.760"", ""start_sec"": 1919.76, ""end_sec"": 1924.32, ""text"": ""square now there are different ways of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:02.159"", ""start_sec"": 1922.159, ""end_sec"": 1926.24, ""text"": ""defining cost function it need not"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:04.320"", ""start_sec"": 1924.32, ""end_sec"": 1928.72, ""text"": ""always be the mean square error but the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:06.240"", ""start_sec"": 1926.24, ""end_sec"": 1931.76, ""text"": ""most common one is this so you define"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:08.720"", ""start_sec"": 1928.72, ""end_sec"": 1935.2, ""text"": ""this cost function and you ask the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:11.760"", ""start_sec"": 1931.76, ""end_sec"": 1937.6, ""text"": ""system to minimize this error so we use"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:15.200"", ""start_sec"": 1935.2, ""end_sec"": 1941.039, ""text"": ""what is known as an optimization"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:17.600"", ""start_sec"": 1937.6, ""end_sec"": 1943.519, ""text"": ""function to minimize this error and the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:21.039"", ""start_sec"": 1941.039, ""end_sec"": 1945.84, ""text"": ""error itself sent back to the system as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:23.519"", ""start_sec"": 1943.519, ""end_sec"": 1947.6, ""text"": ""feedback and that is known as back"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:25.840"", ""start_sec"": 1945.84, ""end_sec"": 1950.08, ""text"": ""propagation and so this is the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:27.600"", ""start_sec"": 1947.6, ""end_sec"": 1952.88, ""text"": ""function and how do we optimize the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:30.080"", ""start_sec"": 1950.08, ""end_sec"": 1955.919, ""text"": ""function we use what is known as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:32.880"", ""start_sec"": 1952.88, ""end_sec"": 1959.519, ""text"": ""gradient descent so the gradient descent"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:35.919"", ""start_sec"": 1955.919, ""end_sec"": 1962.48, ""text"": ""mechanism identifies how to change the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:39.519"", ""start_sec"": 1959.519, ""end_sec"": 1965.76, ""text"": ""weights and biases so that the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:42.480"", ""start_sec"": 1962.48, ""end_sec"": 1967.919, ""text"": ""function is minimized and there is also"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:45.760"", ""start_sec"": 1965.76, ""end_sec"": 1970.159, ""text"": ""what is known as the rate or the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:47.919"", ""start_sec"": 1967.919, ""end_sec"": 1973.12, ""text"": ""learning rate that is what is shown here"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:50.159"", ""start_sec"": 1970.159, ""end_sec"": 1975.919, ""text"": ""as slower and faster so you need to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:53.120"", ""start_sec"": 1973.12, ""end_sec"": 1979.36, ""text"": ""specify what should be the learning rate"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:32:55.919"", ""start_sec"": 1975.919, ""end_sec"": 1982.48, ""text"": ""now if the learning rate is very small"", ""video_id"": ""FbxTVRfQFuI""}]",1878.399,1982.48
then it will probably take very long totrainwhereas if the learning rate is veryhighthen it will appear to befasterbut then it will probablyneverwhat is known asconvergenow what isconvergencenow we are talking about a few termshereconvergence is likethisthis is a representation ofconvergenceso the whole idea of gradient descent is to optimize the cost function or minimize the costfunctionin order to dothatwe need to represent the cost function as thiscurvewe need to come to this minimumvaluethat is what is known as the minimization of the costfunctionnow what happens if we have the learning rate very small is that it will take very long to come to thispointon the otherhandif you havelargehigher learningratewhat will happenisinstead of stoppinghereit will cross over because the learning rate ishighand then it has to comebackso it will result in what is known as like anoscillationso it will never come to thispointwhich is known asconvergenceinstead it will go back andforthso these are known as hyperparametersthe learning rate and soonand these have to be those numbers or those values we candeterminetypically using trial and error out ofexperiencewe we try to find out thesevaluesso that is the gradient descentmechanismto optimize the costfunctionand that is what is used to train our neuralnetworkthis is another representation of how the training processworksandherein thisexamplewe are trying to classify theseimageswhether they are cats ordogsand,FbxTVRfQFuI,00:32:59.360,"[{""timestamp"": ""00:32:59.360"", ""start_sec"": 1979.36, ""end_sec"": 1984.48, ""text"": ""then it will probably take very long to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:02.480"", ""start_sec"": 1982.48, ""end_sec"": 1987.279, ""text"": ""train whereas if the learning rate is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:04.480"", ""start_sec"": 1984.48, ""end_sec"": 1989.84, ""text"": ""very high then it will appear to be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:07.279"", ""start_sec"": 1987.279, ""end_sec"": 1992.159, ""text"": ""faster but then it will probably never"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:09.840"", ""start_sec"": 1989.84, ""end_sec"": 1994.48, ""text"": ""what is known as converge now what is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:12.159"", ""start_sec"": 1992.159, ""end_sec"": 1997.76, ""text"": ""convergence now we are talking about a"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:14.480"", ""start_sec"": 1994.48, ""end_sec"": 2000.159, ""text"": ""few terms here convergence is like this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:17.760"", ""start_sec"": 1997.76, ""end_sec"": 2004.0, ""text"": ""this is a representation of convergence"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:20.159"", ""start_sec"": 2000.159, ""end_sec"": 2006.24, ""text"": ""so the whole idea of gradient descent is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:24.000"", ""start_sec"": 2004.0, ""end_sec"": 2008.64, ""text"": ""to optimize the cost function or"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:26.240"", ""start_sec"": 2006.24, ""end_sec"": 2010.88, ""text"": ""minimize the cost function in order to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:28.640"", ""start_sec"": 2008.64, ""end_sec"": 2014.0, ""text"": ""do that we need to represent the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:30.880"", ""start_sec"": 2010.88, ""end_sec"": 2016.48, ""text"": ""function as this curve we need to come"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:34.000"", ""start_sec"": 2014.0, ""end_sec"": 2018.96, ""text"": ""to this minimum value that is what is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:36.480"", ""start_sec"": 2016.48, ""end_sec"": 2021.84, ""text"": ""known as the minimization of the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:38.960"", ""start_sec"": 2018.96, ""end_sec"": 2024.72, ""text"": ""function now what happens if we have the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:41.840"", ""start_sec"": 2021.84, ""end_sec"": 2028.0, ""text"": ""learning rate very small is that it will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:44.720"", ""start_sec"": 2024.72, ""end_sec"": 2031.2, ""text"": ""take very long to come to this point on"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:48.000"", ""start_sec"": 2028.0, ""end_sec"": 2033.279, ""text"": ""the other hand if you have large higher"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:51.200"", ""start_sec"": 2031.2, ""end_sec"": 2036.159, ""text"": ""learning rate what will happen is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:53.279"", ""start_sec"": 2033.279, ""end_sec"": 2038.559, ""text"": ""instead of stopping here it will cross"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:56.159"", ""start_sec"": 2036.159, ""end_sec"": 2041.279, ""text"": ""over because the learning rate is high"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:33:58.559"", ""start_sec"": 2038.559, ""end_sec"": 2043.44, ""text"": ""and then it has to come back so it will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:01.279"", ""start_sec"": 2041.279, ""end_sec"": 2045.44, ""text"": ""result in what is known as like an"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:03.440"", ""start_sec"": 2043.44, ""end_sec"": 2047.76, ""text"": ""oscillation so it will never come to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:05.440"", ""start_sec"": 2045.44, ""end_sec"": 2050.639, ""text"": ""this point which is known as convergence"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:07.760"", ""start_sec"": 2047.76, ""end_sec"": 2053.04, ""text"": ""instead it will go back and forth so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:10.639"", ""start_sec"": 2050.639, ""end_sec"": 2054.96, ""text"": ""these are known as hyper parameters the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:13.040"", ""start_sec"": 2053.04, ""end_sec"": 2057.52, ""text"": ""learning rate and so on and these have"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:14.960"", ""start_sec"": 2054.96, ""end_sec"": 2060.639, ""text"": ""to be those numbers or those values we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:17.520"", ""start_sec"": 2057.52, ""end_sec"": 2063.04, ""text"": ""can determine typically using trial and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:20.639"", ""start_sec"": 2060.639, ""end_sec"": 2065.359, ""text"": ""error out of experience we we try to"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:23.040"", ""start_sec"": 2063.04, ""end_sec"": 2068.639, ""text"": ""find out these values so that is the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:25.359"", ""start_sec"": 2065.359, ""end_sec"": 2070.56, ""text"": ""gradient descent mechanism to optimize"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:28.639"", ""start_sec"": 2068.639, ""end_sec"": 2074.399, ""text"": ""the cost function and that is what is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:30.560"", ""start_sec"": 2070.56, ""end_sec"": 2076.56, ""text"": ""used to train our neural network this is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:34.399"", ""start_sec"": 2074.399, ""end_sec"": 2078.72, ""text"": ""another representation of how the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:36.560"", ""start_sec"": 2076.56, ""end_sec"": 2081.2, ""text"": ""training process works and here in this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:38.720"", ""start_sec"": 2078.72, ""end_sec"": 2084.32, ""text"": ""example we are trying to classify these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:41.200"", ""start_sec"": 2081.2, ""end_sec"": 2086.96, ""text"": ""images whether they are cats or dogs and"", ""video_id"": ""FbxTVRfQFuI""}]",1979.36,2086.96
as you canseeactually each image is fed in eachtimeone image isfed ratherand these values ofx1 x2up to xn are the pixel values within thisimageokayso those values are then taken and for each of those values a weight is multiplied and then it goes to the nextlayerand then to the nextlayerand soonultimately it comes as the output layer and it gives an outputaswhether it is a dog or acatrememberthe output will never be a namedoutputso these would be like a zero or aoneand wesay okayzero corresponds to dogs and one corresponds tocatchso that is the way it typicallyhappensthis is a binaryclassificationwe have similar situations where there can be multipleclasseswhich means that there will be multiple more neurons in the outputlayerokayso this is once again a quick representation of how the forward propagation and the backward propagationworksso the information is going in thisdirectionwhich is basically the forwardpropagationand at the output level we find out what is the costfunctionthe difference is basically sent back as part of the backward propagation and gradientdescentthen adjust the weights and biases for the nextiterationthis happens iteratively till the cost function isminimizedand that is when wesaythewholethe network hasconvergedor the training process hasconvergedand there can be situations where convergence may nothappenin rarecasesbut by and large the network will converge and after maybe afew,FbxTVRfQFuI,00:34:44.320,"[{""timestamp"": ""00:34:44.320"", ""start_sec"": 2084.32, ""end_sec"": 2089.599, ""text"": ""as you can see actually each image is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:46.960"", ""start_sec"": 2086.96, ""end_sec"": 2094.0, ""text"": ""fed in each time one image is fed rather"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:49.599"", ""start_sec"": 2089.599, ""end_sec"": 2096.96, ""text"": ""and these values of x1 x2 up to xn are"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:54.000"", ""start_sec"": 2094.0, ""end_sec"": 2099.28, ""text"": ""the pixel values within this image okay"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:56.960"", ""start_sec"": 2096.96, ""end_sec"": 2101.92, ""text"": ""so those values are then taken and for"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:34:59.280"", ""start_sec"": 2099.28, ""end_sec"": 2104.32, ""text"": ""each of those values a weight is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:01.920"", ""start_sec"": 2101.92, ""end_sec"": 2106.079, ""text"": ""multiplied and then it goes to the next"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:04.320"", ""start_sec"": 2104.32, ""end_sec"": 2108.48, ""text"": ""layer and then to the next layer and so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:06.079"", ""start_sec"": 2106.079, ""end_sec"": 2110.88, ""text"": ""on ultimately it comes as the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:08.480"", ""start_sec"": 2108.48, ""end_sec"": 2113.839, ""text"": ""layer and it gives an output as whether"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:10.880"", ""start_sec"": 2110.88, ""end_sec"": 2116.72, ""text"": ""it is a dog or a cat remember the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:13.839"", ""start_sec"": 2113.839, ""end_sec"": 2119.52, ""text"": ""will never be a named output so these"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:16.720"", ""start_sec"": 2116.72, ""end_sec"": 2122.4, ""text"": ""would be like a zero or a one and we say"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:19.520"", ""start_sec"": 2119.52, ""end_sec"": 2124.4, ""text"": ""okay zero corresponds to dogs and one"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:22.400"", ""start_sec"": 2122.4, ""end_sec"": 2126.64, ""text"": ""corresponds to catch so that is the way"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:24.400"", ""start_sec"": 2124.4, ""end_sec"": 2128.8, ""text"": ""it typically happens this is a binary"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:26.640"", ""start_sec"": 2126.64, ""end_sec"": 2131.28, ""text"": ""classification we have similar"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:28.800"", ""start_sec"": 2128.8, ""end_sec"": 2132.96, ""text"": ""situations where there can be multiple"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:31.280"", ""start_sec"": 2131.28, ""end_sec"": 2134.96, ""text"": ""classes which means that there will be"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:32.960"", ""start_sec"": 2132.96, ""end_sec"": 2138.16, ""text"": ""multiple more neurons in the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:34.960"", ""start_sec"": 2134.96, ""end_sec"": 2139.92, ""text"": ""layer okay so this is once again a quick"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:38.160"", ""start_sec"": 2138.16, ""end_sec"": 2141.839, ""text"": ""representation of how the forward"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:39.920"", ""start_sec"": 2139.92, ""end_sec"": 2144.4, ""text"": ""propagation and the backward propagation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:41.839"", ""start_sec"": 2141.839, ""end_sec"": 2146.64, ""text"": ""works so the information is going"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:44.400"", ""start_sec"": 2144.4, ""end_sec"": 2149.119, ""text"": ""in this direction which is basically the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:46.640"", ""start_sec"": 2146.64, ""end_sec"": 2150.079, ""text"": ""forward propagation and at the output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:49.119"", ""start_sec"": 2149.119, ""end_sec"": 2153.2, ""text"": ""level"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:50.079"", ""start_sec"": 2150.079, ""end_sec"": 2156.48, ""text"": ""we find out what is the cost function"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:53.200"", ""start_sec"": 2153.2, ""end_sec"": 2158.56, ""text"": ""the difference is basically sent back as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:56.480"", ""start_sec"": 2156.48, ""end_sec"": 2161.04, ""text"": ""part of the backward propagation and"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:35:58.560"", ""start_sec"": 2158.56, ""end_sec"": 2163.52, ""text"": ""gradient descent then adjust the weights"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:01.040"", ""start_sec"": 2161.04, ""end_sec"": 2166.16, ""text"": ""and biases for the next iteration this"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:03.520"", ""start_sec"": 2163.52, ""end_sec"": 2169.28, ""text"": ""happens iteratively till the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:06.160"", ""start_sec"": 2166.16, ""end_sec"": 2171.68, ""text"": ""function is minimized and that is when"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:09.280"", ""start_sec"": 2169.28, ""end_sec"": 2173.76, ""text"": ""we say the whole the network has"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:11.680"", ""start_sec"": 2171.68, ""end_sec"": 2176.16, ""text"": ""converged or the training process has"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:13.760"", ""start_sec"": 2173.76, ""end_sec"": 2178.88, ""text"": ""converged and there can be situations"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:16.160"", ""start_sec"": 2176.16, ""end_sec"": 2181.599, ""text"": ""where convergence may not happen in rare"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:18.880"", ""start_sec"": 2178.88, ""end_sec"": 2184.0, ""text"": ""cases but by and large the network will"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:21.599"", ""start_sec"": 2181.599, ""end_sec"": 2186.32, ""text"": ""converge and after maybe a few"", ""video_id"": ""FbxTVRfQFuI""}]",2084.32,2186.32
iterationsit could be tens of iterations or hundreds ofiterationsdepending on whatexactlythe number of iterations canvaryand then wesay okaywe are getting a certain accuracy and we say that is ourthresholdmaybe 90accuracywe stop at that and we say that the system istrainedthe trained model is then deployed for production and soonso that is the way the neural network traininghappensokayso that is the way classification works in deep learning using neuralnetworkand this slide is an animation of this wholeprocessas you canseethe forwardpropagationthe data is going forward from the input layer to the output layer and there is an output and the error iscalculatedthe cost function is calculated and that is fed back as a part of backward propagation and that whole process repeats onceagainokaysorememberin neural networks the training process is nothing but the finding the best values of the weights and biases for each and every neuron in thenetworkthat'salltraining of neural network consists of finding the optimal values of the weights and biases so that the accuracy ismaximumallrightso with that we come to the end of thesessionwe all have a greatdaythank you verymuchhithereif you like thisvideosubscribe to the simply learn youtube channel and click here to watch similarvideosturn it up and getcertifiedclickhere,FbxTVRfQFuI,00:36:24.000,"[{""timestamp"": ""00:36:24.000"", ""start_sec"": 2184.0, ""end_sec"": 2188.16, ""text"": ""iterations it could be tens of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:26.320"", ""start_sec"": 2186.32, ""end_sec"": 2190.16, ""text"": ""iterations or hundreds of iterations"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:28.160"", ""start_sec"": 2188.16, ""end_sec"": 2192.8, ""text"": ""depending on what exactly the number of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:30.160"", ""start_sec"": 2190.16, ""end_sec"": 2195.599, ""text"": ""iterations can vary and then we say okay"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:32.800"", ""start_sec"": 2192.8, ""end_sec"": 2198.079, ""text"": ""we are getting a certain accuracy and we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:35.599"", ""start_sec"": 2195.599, ""end_sec"": 2200.8, ""text"": ""say that is our threshold maybe 90"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:38.079"", ""start_sec"": 2198.079, ""end_sec"": 2202.88, ""text"": ""accuracy we stop at that and we say that"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:40.800"", ""start_sec"": 2200.8, ""end_sec"": 2204.64, ""text"": ""the system is trained the trained model"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:42.880"", ""start_sec"": 2202.88, ""end_sec"": 2207.44, ""text"": ""is then deployed for production and so"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:44.640"", ""start_sec"": 2204.64, ""end_sec"": 2209.92, ""text"": ""on so that is the way the neural network"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:47.440"", ""start_sec"": 2207.44, ""end_sec"": 2213.2, ""text"": ""training happens okay so that is the way"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:49.920"", ""start_sec"": 2209.92, ""end_sec"": 2216.079, ""text"": ""classification works in deep learning"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:53.200"", ""start_sec"": 2213.2, ""end_sec"": 2219.28, ""text"": ""using neural network and this slide is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:56.079"", ""start_sec"": 2216.079, ""end_sec"": 2221.52, ""text"": ""an animation of this whole process as"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:36:59.280"", ""start_sec"": 2219.28, ""end_sec"": 2224.079, ""text"": ""you can see the forward propagation the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:01.520"", ""start_sec"": 2221.52, ""end_sec"": 2226.16, ""text"": ""data is going forward from the input"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:04.079"", ""start_sec"": 2224.079, ""end_sec"": 2227.359, ""text"": ""layer to the output layer and there is"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:06.160"", ""start_sec"": 2226.16, ""end_sec"": 2230.0, ""text"": ""an output"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:07.359"", ""start_sec"": 2227.359, ""end_sec"": 2232.96, ""text"": ""and the error is calculated the cost"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:10.000"", ""start_sec"": 2230.0, ""end_sec"": 2235.359, ""text"": ""function is calculated and that is fed"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:12.960"", ""start_sec"": 2232.96, ""end_sec"": 2238.32, ""text"": ""back as a part of backward propagation"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:15.359"", ""start_sec"": 2235.359, ""end_sec"": 2240.8, ""text"": ""and that whole process repeats once"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:18.320"", ""start_sec"": 2238.32, ""end_sec"": 2243.359, ""text"": ""again okay so remember in neural"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:20.800"", ""start_sec"": 2240.8, ""end_sec"": 2247.52, ""text"": ""networks the training process is nothing"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:23.359"", ""start_sec"": 2243.359, ""end_sec"": 2249.76, ""text"": ""but the finding the best values of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:27.520"", ""start_sec"": 2247.52, ""end_sec"": 2252.4, ""text"": ""weights and biases for each and every"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:29.760"", ""start_sec"": 2249.76, ""end_sec"": 2254.96, ""text"": ""neuron in the network that's all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:32.400"", ""start_sec"": 2252.4, ""end_sec"": 2257.76, ""text"": ""training of neural network consists of"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:34.960"", ""start_sec"": 2254.96, ""end_sec"": 2260.96, ""text"": ""finding the optimal values of the"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:37.760"", ""start_sec"": 2257.76, ""end_sec"": 2264.8, ""text"": ""weights and biases so that the accuracy"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:40.960"", ""start_sec"": 2260.96, ""end_sec"": 2267.04, ""text"": ""is maximum all right so with that we"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:44.800"", ""start_sec"": 2264.8, ""end_sec"": 2271.8, ""text"": ""come to the end of the session we all"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:47.040"", ""start_sec"": 2267.04, ""end_sec"": 2271.8, ""text"": ""have a great day thank you very much"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:53.839"", ""start_sec"": 2273.839, ""end_sec"": 2277.359, ""text"": ""hi there if you like this video"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:55.680"", ""start_sec"": 2275.68, ""end_sec"": 2280.0, ""text"": ""subscribe to the simply learn youtube"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:37:57.359"", ""start_sec"": 2277.359, ""end_sec"": 2282.16, ""text"": ""channel and click here to watch similar"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:38:00.000"", ""start_sec"": 2280.0, ""end_sec"": 2285.48, ""text"": ""videos turn it up and get certified"", ""video_id"": ""FbxTVRfQFuI""}, {""timestamp"": ""00:38:02.160"", ""start_sec"": 2282.16, ""end_sec"": 2285.48, ""text"": ""click here"", ""video_id"": ""FbxTVRfQFuI""}]",2184.0,2285.48
weknowhumans learn from their past experiences and machines follow instructions given byhumansbut what if humans can train the machines to learn from the past data and do what humans candoand muchfasterwellthat's called machinelearningbut it's a lot more than justlearningit's also about understanding andreasoningso today we will learn about the basics of machinelearningso that'spaulhe loves listening to newsongshe either likes them or dislikesthempaul decides this on the basis of the song'stempo genreintensity and the gender ofvoiceforsimplicitylet's just use tempo and intensity fornowso here tempo is on the xaxisranging from relaxed tofastwhereas intensity is on the yaxisranging from light tosoaringwe see that paul likes the song with fast tempo and soaringintensitywhile he dislikes the song with relaxed tempo and lightintensityso now we know paul'schoiceslet's say paul listens to a newsonglet's name it as songasong a has fast tempo and a soaringintensityso it lies somewhereherelooking at thedatacan you guess whether paul will like the song ornot correctso paul likes thissongby looking at paul's pastchoiceswe were able to classify the unknown song veryeasilyrightlet's say now paul listens to a newsonglet's label it as songbso song b lies somewhereherewith mediumtempo,ukzFI9rgwfU,00:00:00.240,"[{""timestamp"": ""00:00:00.240"", ""start_sec"": 0.24, ""end_sec"": 3.76, ""text"": ""we know humans learn from their past"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:02.320"", ""start_sec"": 2.32, ""end_sec"": 5.68, ""text"": ""experiences"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:03.760"", ""start_sec"": 3.76, ""end_sec"": 7.359, ""text"": ""and machines follow instructions given"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:05.680"", ""start_sec"": 5.68, ""end_sec"": 9.599, ""text"": ""by humans"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:07.359"", ""start_sec"": 7.359, ""end_sec"": 11.519, ""text"": ""but what if humans can train the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:09.599"", ""start_sec"": 9.599, ""end_sec"": 14.0, ""text"": ""machines to learn from the past data and"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:11.519"", ""start_sec"": 11.519, ""end_sec"": 15.839, ""text"": ""do what humans can do and much faster"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:14.000"", ""start_sec"": 14.0, ""end_sec"": 17.76, ""text"": ""well that's called machine learning but"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:15.839"", ""start_sec"": 15.839, ""end_sec"": 20.0, ""text"": ""it's a lot more than just learning it's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:17.760"", ""start_sec"": 17.76, ""end_sec"": 22.4, ""text"": ""also about understanding and reasoning"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:20.000"", ""start_sec"": 20.0, ""end_sec"": 24.24, ""text"": ""so today we will learn about the basics"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:22.400"", ""start_sec"": 22.4, ""end_sec"": 26.8, ""text"": ""of machine learning"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:24.240"", ""start_sec"": 24.24, ""end_sec"": 28.8, ""text"": ""so that's paul he loves listening to new"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:26.800"", ""start_sec"": 26.8, ""end_sec"": 30.88, ""text"": ""songs"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:28.800"", ""start_sec"": 28.8, ""end_sec"": 33.12, ""text"": ""he either likes them or dislikes them"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:30.880"", ""start_sec"": 30.88, ""end_sec"": 34.88, ""text"": ""paul decides this on the basis of the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:33.120"", ""start_sec"": 33.12, ""end_sec"": 36.0, ""text"": ""song's tempo"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:34.880"", ""start_sec"": 34.88, ""end_sec"": 39.04, ""text"": ""genre"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:36.000"", ""start_sec"": 36.0, ""end_sec"": 41.44, ""text"": ""intensity and the gender of voice for"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:39.040"", ""start_sec"": 39.04, ""end_sec"": 44.559, ""text"": ""simplicity let's just use tempo and"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:41.440"", ""start_sec"": 41.44, ""end_sec"": 47.68, ""text"": ""intensity for now so here tempo is on"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:44.559"", ""start_sec"": 44.559, ""end_sec"": 50.32, ""text"": ""the x axis ranging from relaxed to fast"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:47.680"", ""start_sec"": 47.68, ""end_sec"": 53.28, ""text"": ""whereas intensity is on the y axis"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:50.320"", ""start_sec"": 50.32, ""end_sec"": 56.879, ""text"": ""ranging from light to soaring we see"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:53.280"", ""start_sec"": 53.28, ""end_sec"": 59.84, ""text"": ""that paul likes the song with fast tempo"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:56.879"", ""start_sec"": 56.879, ""end_sec"": 62.8, ""text"": ""and soaring intensity while he dislikes"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:00:59.840"", ""start_sec"": 59.84, ""end_sec"": 65.28, ""text"": ""the song with relaxed tempo and light"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:02.800"", ""start_sec"": 62.8, ""end_sec"": 67.36, ""text"": ""intensity so now we know paul's choices"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:05.280"", ""start_sec"": 65.28, ""end_sec"": 70.72, ""text"": ""let's say paul listens to a new song"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:07.360"", ""start_sec"": 67.36, ""end_sec"": 73.68, ""text"": ""let's name it as song a song a has fast"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:10.720"", ""start_sec"": 70.72, ""end_sec"": 75.84, ""text"": ""tempo and a soaring intensity so it lies"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:13.680"", ""start_sec"": 73.68, ""end_sec"": 77.759, ""text"": ""somewhere here looking at the data can"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:15.840"", ""start_sec"": 75.84, ""end_sec"": 80.56, ""text"": ""you guess whether paul will like the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:17.759"", ""start_sec"": 77.759, ""end_sec"": 83.03999999999999, ""text"": ""song or not correct so paul likes this"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:20.560"", ""start_sec"": 80.56, ""end_sec"": 85.119, ""text"": ""song by looking at paul's past choices"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:23.040"", ""start_sec"": 83.03999999999999, ""end_sec"": 88.4, ""text"": ""we were able to classify the unknown"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:25.119"", ""start_sec"": 85.119, ""end_sec"": 90.88, ""text"": ""song very easily right let's say now"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:28.400"", ""start_sec"": 88.4, ""end_sec"": 93.439, ""text"": ""paul listens to a new song let's label"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:30.880"", ""start_sec"": 90.88, ""end_sec"": 96.72, ""text"": ""it as song b so song b"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:33.439"", ""start_sec"": 93.439, ""end_sec"": 99.439, ""text"": ""lies somewhere here with medium tempo"", ""video_id"": ""ukzFI9rgwfU""}]",0.24,99.439
and mediumintensityneither relaxed norfastneither light norsoaringnow can you guess whether paul likes it ornotnot able to guess whether paul will like it or dislikeitare the choicesunclearcorrectwe could easily classify songabut when the choice becamecomplicatedas in the case of songb yesand that's where machine learning comesinlet's seehowin the same example for songbif we draw a circle around the songbwe see that there are four votes for like whereas one would fordislikeif we go for the majorityvoteswe can say that paul will definitely like thesongthat'sallthis was a basic machine learningalgorithmalsoit's called k nearestneighborsso this is just a small example in one of the many machine learningalgorithmsquiteeasy rightbelieve me itisbut what happens when the choices becomecomplicatedas in the case of songbthat's when machine learning comesinit learns thedatabuilds the prediction model and when the new data point comesinit can easily predict foritmore thedatabetter themodelhigher will be theaccuracythere are many ways in which the machinelearnsit could be either supervisedlearningunsupervised learning or reinforcementlearninglet's first quickly understand supervisedlearningsuppose your friend gives you one million coins of three differentcurrenciessay onerupeeone euro and onedirhameach coin has differentweightsforexamplea coin of one rupee weighs threegramsone euro weighs seven grams and one dirham weighs fourgrams,ukzFI9rgwfU,00:01:36.720,"[{""timestamp"": ""00:01:36.720"", ""start_sec"": 96.72, ""end_sec"": 102.4, ""text"": ""and medium intensity neither relaxed nor"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:39.439"", ""start_sec"": 99.439, ""end_sec"": 104.479, ""text"": ""fast neither light nor soaring now can"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:42.400"", ""start_sec"": 102.4, ""end_sec"": 106.56, ""text"": ""you guess whether paul likes it or not"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:44.479"", ""start_sec"": 104.479, ""end_sec"": 109.2, ""text"": ""not able to guess whether paul will like"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:46.560"", ""start_sec"": 106.56, ""end_sec"": 112.15899999999999, ""text"": ""it or dislike it are the choices unclear"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:49.200"", ""start_sec"": 109.2, ""end_sec"": 114.64, ""text"": ""correct we could easily classify song a"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:52.159"", ""start_sec"": 112.15899999999999, ""end_sec"": 117.2, ""text"": ""but when the choice became complicated"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:54.640"", ""start_sec"": 114.64, ""end_sec"": 119.119, ""text"": ""as in the case of song b yes and that's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:57.200"", ""start_sec"": 117.2, ""end_sec"": 121.92, ""text"": ""where machine learning comes in let's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:01:59.119"", ""start_sec"": 119.119, ""end_sec"": 124.24, ""text"": ""see how in the same example for song b"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:01.920"", ""start_sec"": 121.92, ""end_sec"": 126.719, ""text"": ""if we draw a circle around the song b we"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:04.240"", ""start_sec"": 124.24, ""end_sec"": 129.44, ""text"": ""see that there are four votes for like"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:06.719"", ""start_sec"": 126.719, ""end_sec"": 131.76, ""text"": ""whereas one would for dislike if we go"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:09.440"", ""start_sec"": 129.44, ""end_sec"": 133.44, ""text"": ""for the majority votes we can say that"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:11.760"", ""start_sec"": 131.76, ""end_sec"": 135.12, ""text"": ""paul will definitely like the song"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:13.440"", ""start_sec"": 133.44, ""end_sec"": 137.12, ""text"": ""that's all this was a basic machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:15.120"", ""start_sec"": 135.12, ""end_sec"": 139.2, ""text"": ""learning algorithm also it's called k"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:17.120"", ""start_sec"": 137.12, ""end_sec"": 141.599, ""text"": ""nearest neighbors so this is just a"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:19.200"", ""start_sec"": 139.2, ""end_sec"": 144.319, ""text"": ""small example in one of the many machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:21.599"", ""start_sec"": 141.599, ""end_sec"": 147.44, ""text"": ""learning algorithms quite easy right"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:24.319"", ""start_sec"": 144.319, ""end_sec"": 149.84, ""text"": ""believe me it is but what happens when"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:27.440"", ""start_sec"": 147.44, ""end_sec"": 151.76, ""text"": ""the choices become complicated as in the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:29.840"", ""start_sec"": 149.84, ""end_sec"": 153.92000000000002, ""text"": ""case of song b that's when machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:31.760"", ""start_sec"": 151.76, ""end_sec"": 155.92000000000002, ""text"": ""learning comes in it learns the data"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:33.920"", ""start_sec"": 153.92000000000002, ""end_sec"": 158.16, ""text"": ""builds the prediction model and when the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:35.920"", ""start_sec"": 155.92000000000002, ""end_sec"": 160.64, ""text"": ""new data point comes in it can easily"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:38.160"", ""start_sec"": 158.16, ""end_sec"": 163.2, ""text"": ""predict for it more the data better the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:40.640"", ""start_sec"": 160.64, ""end_sec"": 165.36, ""text"": ""model higher will be the accuracy there"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:43.200"", ""start_sec"": 163.2, ""end_sec"": 167.599, ""text"": ""are many ways in which the machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:45.360"", ""start_sec"": 165.36, ""end_sec"": 169.599, ""text"": ""learns it could be either supervised"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:47.599"", ""start_sec"": 167.599, ""end_sec"": 171.28, ""text"": ""learning unsupervised learning or"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:49.599"", ""start_sec"": 169.599, ""end_sec"": 173.68, ""text"": ""reinforcement learning let's first"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:51.280"", ""start_sec"": 171.28, ""end_sec"": 175.519, ""text"": ""quickly understand supervised learning"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:53.680"", ""start_sec"": 173.68, ""end_sec"": 177.28, ""text"": ""suppose your friend gives you one"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:55.519"", ""start_sec"": 175.519, ""end_sec"": 180.0, ""text"": ""million coins of three different"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:02:57.280"", ""start_sec"": 177.28, ""end_sec"": 182.08, ""text"": ""currencies say one rupee one euro and"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:00.000"", ""start_sec"": 180.0, ""end_sec"": 184.48, ""text"": ""one dirham each coin has different"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:02.080"", ""start_sec"": 182.08, ""end_sec"": 187.12, ""text"": ""weights for example a coin of one rupee"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:04.480"", ""start_sec"": 184.48, ""end_sec"": 189.519, ""text"": ""weighs three grams one euro weighs seven"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:07.120"", ""start_sec"": 187.12, ""end_sec"": 191.44, ""text"": ""grams and one dirham weighs four grams"", ""video_id"": ""ukzFI9rgwfU""}]",96.72,191.44
your model will predict the currency of thecoinhere your weight becomes the feature ofcoinswhile currency becomes thelabelwhen you feed this data to the machine learningmodelit learns which feature is associated with whichlabelforexampleit will learn that if a coin is of 3 grams it will be a 1 rupeecoinlet's give a new coin to themachineon the basis of the weight of the newcoinyour model will predict thecurrencyhence supervised learning uses labeled data to train themodelhere the machine knew the features of the object and also the labels associated with thosefeatureson thisnotelet's move to unsupervised learning and see thedifferencesuppose you have cricket data set of various players with their respective scores and wicketstakenwhen you feed this data set to themachinethe machine identifies the pattern of playerperformanceso it plots this data with the respective wickets on the x-axis while runs on they-axiswhile looking at thedatayou'll clearly see that there are twoclustersthe one cluster are the players who scored higher runs and took lesswicketswhile the other cluster is of the players who scored less runs but took manywicketsso here we interpret these two clusters as batsmen andbowlersthe important point to note here is that there were no labels of batsmen andbowlershence the learning with unlabeled data is unsupervisedlearningso we saw supervisedlearningwhere the data waslabeledand the unsupervisedlearningwhere the data was unlabeled and then,ukzFI9rgwfU,00:03:09.519,"[{""timestamp"": ""00:03:09.519"", ""start_sec"": 189.519, ""end_sec"": 193.92, ""text"": ""your model will predict the currency of"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:11.440"", ""start_sec"": 191.44, ""end_sec"": 196.4, ""text"": ""the coin here your weight becomes the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:13.920"", ""start_sec"": 193.92, ""end_sec"": 198.4, ""text"": ""feature of coins while currency becomes"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:16.400"", ""start_sec"": 196.4, ""end_sec"": 201.04, ""text"": ""the label when you feed this data to the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:18.400"", ""start_sec"": 198.4, ""end_sec"": 203.68, ""text"": ""machine learning model it learns which"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:21.040"", ""start_sec"": 201.04, ""end_sec"": 206.319, ""text"": ""feature is associated with which label"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:23.680"", ""start_sec"": 203.68, ""end_sec"": 208.959, ""text"": ""for example it will learn that if a coin"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:26.319"", ""start_sec"": 206.319, ""end_sec"": 210.56, ""text"": ""is of 3 grams it will be a 1 rupee coin"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:28.959"", ""start_sec"": 208.959, ""end_sec"": 212.879, ""text"": ""let's give a new coin to the machine on"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:30.560"", ""start_sec"": 210.56, ""end_sec"": 214.959, ""text"": ""the basis of the weight of the new coin"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:32.879"", ""start_sec"": 212.879, ""end_sec"": 217.599, ""text"": ""your model will predict the currency"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:34.959"", ""start_sec"": 214.959, ""end_sec"": 220.0, ""text"": ""hence supervised learning uses labeled"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:37.599"", ""start_sec"": 217.599, ""end_sec"": 222.4, ""text"": ""data to train the model here the machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:40.000"", ""start_sec"": 220.0, ""end_sec"": 224.159, ""text"": ""knew the features of the object and also"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:42.400"", ""start_sec"": 222.4, ""end_sec"": 226.159, ""text"": ""the labels associated with those"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:44.159"", ""start_sec"": 224.159, ""end_sec"": 227.76, ""text"": ""features on this note let's move to"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:46.159"", ""start_sec"": 226.159, ""end_sec"": 229.76, ""text"": ""unsupervised learning and see the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:47.760"", ""start_sec"": 227.76, ""end_sec"": 231.44, ""text"": ""difference suppose you have cricket data"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:49.760"", ""start_sec"": 229.76, ""end_sec"": 233.76, ""text"": ""set of various players with their"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:51.440"", ""start_sec"": 231.44, ""end_sec"": 236.31900000000002, ""text"": ""respective scores and wickets taken when"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:53.760"", ""start_sec"": 233.76, ""end_sec"": 238.64, ""text"": ""you feed this data set to the machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:56.319"", ""start_sec"": 236.31900000000002, ""end_sec"": 240.959, ""text"": ""the machine identifies the pattern of"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:03:58.640"", ""start_sec"": 238.64, ""end_sec"": 242.319, ""text"": ""player performance so it plots this data"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:00.959"", ""start_sec"": 240.959, ""end_sec"": 244.799, ""text"": ""with the respective wickets on the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:02.319"", ""start_sec"": 242.319, ""end_sec"": 246.799, ""text"": ""x-axis while runs on the y-axis while"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:04.799"", ""start_sec"": 244.799, ""end_sec"": 248.879, ""text"": ""looking at the data you'll clearly see"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:06.799"", ""start_sec"": 246.799, ""end_sec"": 250.879, ""text"": ""that there are two clusters the one"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:08.879"", ""start_sec"": 248.879, ""end_sec"": 253.28, ""text"": ""cluster are the players who scored"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:10.879"", ""start_sec"": 250.879, ""end_sec"": 255.68, ""text"": ""higher runs and took less wickets while"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:13.280"", ""start_sec"": 253.28, ""end_sec"": 258.0, ""text"": ""the other cluster is of the players who"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:15.680"", ""start_sec"": 255.68, ""end_sec"": 260.56, ""text"": ""scored less runs but took many wickets"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:18.000"", ""start_sec"": 258.0, ""end_sec"": 262.8, ""text"": ""so here we interpret these two clusters"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:20.560"", ""start_sec"": 260.56, ""end_sec"": 264.8, ""text"": ""as batsmen and bowlers the important"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:22.800"", ""start_sec"": 262.8, ""end_sec"": 267.52, ""text"": ""point to note here is that there were no"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:24.800"", ""start_sec"": 264.8, ""end_sec"": 269.759, ""text"": ""labels of batsmen and bowlers hence the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:27.520"", ""start_sec"": 267.52, ""end_sec"": 271.36, ""text"": ""learning with unlabeled data is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:29.759"", ""start_sec"": 269.759, ""end_sec"": 273.199, ""text"": ""unsupervised learning so we saw"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:31.360"", ""start_sec"": 271.36, ""end_sec"": 275.199, ""text"": ""supervised learning where the data was"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:33.199"", ""start_sec"": 273.199, ""end_sec"": 277.52, ""text"": ""labeled and the unsupervised learning"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:35.199"", ""start_sec"": 275.199, ""end_sec"": 279.36, ""text"": ""where the data was unlabeled and then"", ""video_id"": ""ukzFI9rgwfU""}]",189.519,279.36
there is reinforcementlearningwhich is a reward basedlearningor we can say that it works on the principle offeedbackherelet's say you provide the system with an image of a dog and ask it to identifyitthe system identifies it as acatso you give a negative feedback to themachinesaying that it's a dog'simagethe machine will learn from the feedback andfinallyif it comes across any other image of adogit will be able to classify itcorrectlythat is reinforcementlearningto generalize machine learningmodellet's see aflowchartinput is given to a machine learning model which then gives the output according to the algorithmappliedif it'srightwe take the output as a finalresultelse we provide feedback to the training model and ask it to predict until itlearnsi hope you've understood supervised and unsupervisedlearningso let's have a quickquizyou have to determine whether the given scenarios uses supervised or unsupervisedlearningsimplerightscenarioonefacebook recognizes your friend in a picture from an album of taggedphotographsscenario2netflix recommends new movies based on someone's past moviechoicesscenario3analyzing bank data for suspicious transactions and flagging the fraudtransactionsthink wisely and comment below youranswersmovingondon't you sometimes wonder how is machine learning possible in today'serawellthat's because today we have humongous dataavailableeverybody isonlineeither making a transaction or,ukzFI9rgwfU,00:04:37.520,"[{""timestamp"": ""00:04:37.520"", ""start_sec"": 277.52, ""end_sec"": 281.28, ""text"": ""there is reinforcement learning which is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:39.360"", ""start_sec"": 279.36, ""end_sec"": 282.56, ""text"": ""a reward based learning or we can say"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:41.280"", ""start_sec"": 281.28, ""end_sec"": 284.639, ""text"": ""that it works on the principle of"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:42.560"", ""start_sec"": 282.56, ""end_sec"": 286.96, ""text"": ""feedback here let's say you provide the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:44.639"", ""start_sec"": 284.639, ""end_sec"": 289.919, ""text"": ""system with an image of a dog and ask it"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:46.960"", ""start_sec"": 286.96, ""end_sec"": 292.08, ""text"": ""to identify it the system identifies it"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:49.919"", ""start_sec"": 289.919, ""end_sec"": 294.0, ""text"": ""as a cat so you give a negative feedback"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:52.080"", ""start_sec"": 292.08, ""end_sec"": 295.6, ""text"": ""to the machine saying that it's a dog's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:54.000"", ""start_sec"": 294.0, ""end_sec"": 297.759, ""text"": ""image the machine will learn from the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:55.600"", ""start_sec"": 295.6, ""end_sec"": 299.919, ""text"": ""feedback and finally if it comes across"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:57.759"", ""start_sec"": 297.759, ""end_sec"": 301.919, ""text"": ""any other image of a dog it will be able"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:04:59.919"", ""start_sec"": 299.919, ""end_sec"": 303.84, ""text"": ""to classify it correctly that is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:01.919"", ""start_sec"": 301.919, ""end_sec"": 305.52, ""text"": ""reinforcement learning to generalize"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:03.840"", ""start_sec"": 303.84, ""end_sec"": 307.68, ""text"": ""machine learning model let's see a"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:05.520"", ""start_sec"": 305.52, ""end_sec"": 309.28, ""text"": ""flowchart input is given to a machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:07.680"", ""start_sec"": 307.68, ""end_sec"": 310.96, ""text"": ""learning model which then gives the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:09.280"", ""start_sec"": 309.28, ""end_sec"": 313.52, ""text"": ""output according to the algorithm"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:10.960"", ""start_sec"": 310.96, ""end_sec"": 316.0, ""text"": ""applied if it's right we take the output"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:13.520"", ""start_sec"": 313.52, ""end_sec"": 318.08, ""text"": ""as a final result else we provide"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:16.000"", ""start_sec"": 316.0, ""end_sec"": 320.639, ""text"": ""feedback to the training model and ask"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:18.080"", ""start_sec"": 318.08, ""end_sec"": 322.16, ""text"": ""it to predict until it learns i hope"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:20.639"", ""start_sec"": 320.639, ""end_sec"": 323.919, ""text"": ""you've understood supervised and"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:22.160"", ""start_sec"": 322.16, ""end_sec"": 326.24, ""text"": ""unsupervised learning so let's have a"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:23.919"", ""start_sec"": 323.919, ""end_sec"": 328.72, ""text"": ""quick quiz you have to determine whether"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:26.240"", ""start_sec"": 326.24, ""end_sec"": 330.56, ""text"": ""the given scenarios uses supervised or"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:28.720"", ""start_sec"": 328.72, ""end_sec"": 332.88, ""text"": ""unsupervised learning simple right"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:30.560"", ""start_sec"": 330.56, ""end_sec"": 335.039, ""text"": ""scenario one facebook recognizes your"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:32.880"", ""start_sec"": 332.88, ""end_sec"": 337.52, ""text"": ""friend in a picture from an album of"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:35.039"", ""start_sec"": 335.039, ""end_sec"": 340.639, ""text"": ""tagged photographs"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:37.520"", ""start_sec"": 337.52, ""end_sec"": 343.84000000000003, ""text"": ""scenario 2 netflix recommends new movies"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:40.639"", ""start_sec"": 340.639, ""end_sec"": 346.4, ""text"": ""based on someone's past movie choices"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:43.840"", ""start_sec"": 343.84000000000003, ""end_sec"": 348.8, ""text"": ""scenario 3 analyzing bank data for"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:46.400"", ""start_sec"": 346.4, ""end_sec"": 351.12, ""text"": ""suspicious transactions and flagging the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:48.800"", ""start_sec"": 348.8, ""end_sec"": 353.36, ""text"": ""fraud transactions think wisely and"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:51.120"", ""start_sec"": 351.12, ""end_sec"": 355.44, ""text"": ""comment below your answers moving on"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:53.360"", ""start_sec"": 353.36, ""end_sec"": 357.68, ""text"": ""don't you sometimes wonder how is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:55.440"", ""start_sec"": 355.44, ""end_sec"": 359.28, ""text"": ""machine learning possible in today's era"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:57.680"", ""start_sec"": 357.68, ""end_sec"": 362.0, ""text"": ""well that's because today we have"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:05:59.280"", ""start_sec"": 359.28, ""end_sec"": 364.479, ""text"": ""humongous data available everybody is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:02.000"", ""start_sec"": 362.0, ""end_sec"": 366.24, ""text"": ""online either making a transaction or"", ""video_id"": ""ukzFI9rgwfU""}]",277.52,366.24
just surfing the internet and that's generating a huge amount of data everyminuteand thatdatamyfriendis the key toanalysisalsothe memory handling capabilities of computers have largelyincreasedwhich helps them to process such huge amount of data at hand without anydelayand yescomputers now have great computationalpowersso there are a lot of applications of machine learning outthereto name afewmachine learning is used inhealthcarewhere diagnostics are predicted for doctor'sreviewthe sentiment analysis that the tech giants are doing on social media is another interesting application of machinelearningfraud detection in the finance sector and also to predict customer churn in the e-commercesectorwhile booking agapyou must have encountered surgepricingoften where it says the fair of your trip has beenupdatedcontinuebookingyes pleasei'm getting late forofficewellthat's an interesting machine learning model which is used by global taxigiantuber andotherswhere they have differential pricing in real time based ondemandthe number of carsavailablebadweatherrushretcso they use the surge pricing model to ensure that those who need a cab can getonealsoit uses predictive modeling to predict where the demand will behighwith the goal that drivers can take care of the demand and search pricing can beminimizedgreat hey sirican you remind me to book a cab at 6 pmtodayoki'll remindyouthanksnoproblemcomment below some,ukzFI9rgwfU,00:06:04.479,"[{""timestamp"": ""00:06:04.479"", ""start_sec"": 364.479, ""end_sec"": 368.56, ""text"": ""just surfing the internet and that's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:06.240"", ""start_sec"": 366.24, ""end_sec"": 370.96, ""text"": ""generating a huge amount of data every"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:08.560"", ""start_sec"": 368.56, ""end_sec"": 373.44, ""text"": ""minute and that data my friend is the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:10.960"", ""start_sec"": 370.96, ""end_sec"": 375.52, ""text"": ""key to analysis also the memory handling"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:13.440"", ""start_sec"": 373.44, ""end_sec"": 377.36, ""text"": ""capabilities of computers have largely"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:15.520"", ""start_sec"": 375.52, ""end_sec"": 380.479, ""text"": ""increased which helps them to process"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:17.360"", ""start_sec"": 377.36, ""end_sec"": 383.28, ""text"": ""such huge amount of data at hand without"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:20.479"", ""start_sec"": 380.479, ""end_sec"": 385.36, ""text"": ""any delay and yes computers now have"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:23.280"", ""start_sec"": 383.28, ""end_sec"": 387.28, ""text"": ""great computational powers so there are"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:25.360"", ""start_sec"": 385.36, ""end_sec"": 389.52, ""text"": ""a lot of applications of machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:27.280"", ""start_sec"": 387.28, ""end_sec"": 391.28, ""text"": ""learning out there to name a few machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:29.520"", ""start_sec"": 389.52, ""end_sec"": 393.44, ""text"": ""learning is used in healthcare where"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:31.280"", ""start_sec"": 391.28, ""end_sec"": 395.44, ""text"": ""diagnostics are predicted for doctor's"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:33.440"", ""start_sec"": 393.44, ""end_sec"": 397.759, ""text"": ""review the sentiment analysis that the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:35.440"", ""start_sec"": 395.44, ""end_sec"": 399.6, ""text"": ""tech giants are doing on social media is"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:37.759"", ""start_sec"": 397.759, ""end_sec"": 401.36, ""text"": ""another interesting application of"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:39.600"", ""start_sec"": 399.6, ""end_sec"": 403.28, ""text"": ""machine learning fraud detection in the"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:41.360"", ""start_sec"": 401.36, ""end_sec"": 405.52, ""text"": ""finance sector and also to predict"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:43.280"", ""start_sec"": 403.28, ""end_sec"": 407.12, ""text"": ""customer churn in the e-commerce sector"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:45.520"", ""start_sec"": 405.52, ""end_sec"": 409.759, ""text"": ""while booking a gap you must have"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:47.120"", ""start_sec"": 407.12, ""end_sec"": 411.52, ""text"": ""encountered surge pricing often where it"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:49.759"", ""start_sec"": 409.759, ""end_sec"": 414.24, ""text"": ""says the fair of your trip has been"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:51.520"", ""start_sec"": 411.52, ""end_sec"": 416.0, ""text"": ""updated continue booking yes please i'm"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:54.240"", ""start_sec"": 414.24, ""end_sec"": 418.15999999999997, ""text"": ""getting late for office"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:56.000"", ""start_sec"": 416.0, ""end_sec"": 420.24, ""text"": ""well that's an interesting machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:06:58.160"", ""start_sec"": 418.15999999999997, ""end_sec"": 422.639, ""text"": ""learning model which is used by global"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:00.240"", ""start_sec"": 420.24, ""end_sec"": 424.639, ""text"": ""taxi giant uber and others where they"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:02.639"", ""start_sec"": 422.639, ""end_sec"": 426.56, ""text"": ""have differential pricing in real time"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:04.639"", ""start_sec"": 424.639, ""end_sec"": 430.0, ""text"": ""based on demand the number of cars"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:06.560"", ""start_sec"": 426.56, ""end_sec"": 432.56, ""text"": ""available bad weather rush r etc so they"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:10.000"", ""start_sec"": 430.0, ""end_sec"": 434.8, ""text"": ""use the surge pricing model to ensure"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:12.560"", ""start_sec"": 432.56, ""end_sec"": 437.28, ""text"": ""that those who need a cab can get one"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:14.800"", ""start_sec"": 434.8, ""end_sec"": 439.599, ""text"": ""also it uses predictive modeling to"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:17.280"", ""start_sec"": 437.28, ""end_sec"": 441.68, ""text"": ""predict where the demand will be high"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:19.599"", ""start_sec"": 439.599, ""end_sec"": 443.759, ""text"": ""with the goal that drivers can take care"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:21.680"", ""start_sec"": 441.68, ""end_sec"": 446.319, ""text"": ""of the demand and search pricing can be"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:23.759"", ""start_sec"": 443.759, ""end_sec"": 449.28, ""text"": ""minimized great hey siri can you remind"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:26.319"", ""start_sec"": 446.319, ""end_sec"": 450.4, ""text"": ""me to book a cab at 6 pm today ok i'll"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:29.280"", ""start_sec"": 449.28, ""end_sec"": 453.12, ""text"": ""remind you"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:30.400"", ""start_sec"": 450.4, ""end_sec"": 455.52, ""text"": ""thanks no problem comment below some"", ""video_id"": ""ukzFI9rgwfU""}]",364.479,455.52
interesting everyday examples around you where machines are learning and doing amazingjobsso that's all for machine learning basics today from mysitekeep watching this space for more interestingvideosuntilthenhappylearning,ukzFI9rgwfU,00:07:33.120,"[{""timestamp"": ""00:07:33.120"", ""start_sec"": 453.12, ""end_sec"": 457.36, ""text"": ""interesting everyday examples around you"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:35.520"", ""start_sec"": 455.52, ""end_sec"": 459.84000000000003, ""text"": ""where machines are learning and doing"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:37.360"", ""start_sec"": 457.36, ""end_sec"": 461.84000000000003, ""text"": ""amazing jobs so that's all for machine"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:39.840"", ""start_sec"": 459.84000000000003, ""end_sec"": 463.68, ""text"": ""learning basics today from my site keep"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:41.840"", ""start_sec"": 461.84000000000003, ""end_sec"": 468.199, ""text"": ""watching this space for more interesting"", ""video_id"": ""ukzFI9rgwfU""}, {""timestamp"": ""00:07:43.680"", ""start_sec"": 463.68, ""end_sec"": 468.199, ""text"": ""videos until then happy learning"", ""video_id"": ""ukzFI9rgwfU""}]",453.12,468.199