id
int64
1
546k
paper_arxiv_id
stringlengths
10
13
path
stringlengths
1
249
caption
stringlengths
9
170k
label
stringlengths
6
192
name
float64
100
2008.03847
fig14.pdf
\caption{The phase diagram of chiral symmetry and confinement in the $N_{f}^{\text{cr}}$ vs $T_{c}$ for various values of $eB$. The $N_{f}^{\text{cr}}$ is obtained from the thermal gradient of the chiral condensate $\partial_T \left\langle \bar{\psi} \psi \right\rangle^{1/3} (N_f,eB,T)$ and of the confining scale $\partial_T \widetilde{\tau}_{\text{IR}}^{-1}(N_f,eB,T)$.}
\label{Fig14}
null
101
2008.03847
fig15.pdf
\caption{Magnetic field dependent coupling of Eq.~(\ref{fceB}).}
\label{Fig15}
null
102
2008.03847
fig16.pdf
\caption{The thermal gradient of the chiral condensate $\partial_T \left\langle \bar{\psi} \psi \right\rangle^{1/3} (N_f=2,eB,T)$ has been plotted as a function of $T$ for various values of $eB$ using the magnetic field dependent coupling of Eq.~(\ref{fceB}). It now exhibits IMC. For higher $T$, chiral symmetry breaking occurs at lower values of $eB$.}
\label{Fig16}
null
103
2008.03847
fig17.pdf
\caption{The phase diagram of chiral symmetry and confinement for $N_{f}^{\text{cr}}$ vs $T_{c}$ at finite $eB$ obtained with an $eB$-dependent coupling of Eq.~(\ref{fceB}). The $N_{f}^{\text{cr}}$ is obtained from the thermal gradient of the chiral condensate $\partial_T \left\langle \bar{\psi} \psi \right\rangle^{1/3} (N_f,eB,T)$ and of the confining scale $\partial_T \widetilde{\tau}_{\text{IR}}^{-1}(N_f,eB,T)$}
\label{Fig17}
null
104
2008.03847
fig18.pdf
\caption{The thermal gradient of the chiral condensate $\partial_T \left\langle \bar{\psi} \psi \right\rangle^{1/3} (N_f=2,eB,T)$ has been plotted as a function of $T$ for various values of $eB$ using the magnetic field and temperature dependent coupling of Eq.~(\ref{fceBT}). It again exhibits IMC.}
\label{Fig18}
null
105
2008.03847
fig19.pdf
\caption{The phase diagram of chiral symmetry and confinement for $N_{f}^{\text{cr}}$ vs $T_{c}$ at finite $eB$ obtained now with an $(eB,T)$-dependent coupling of Eq.~(\ref{fceBT}).}
\label{Fig19}
null
106
2508.00602
images/approach/leaksealer_approach.pdf
\caption{Visual representation of the LeakSealer approach.}
\label{fig:leaksealer}
null
107
2508.00602
images/experiments/Static/static_openai.pdf
\caption{OpenAI dataset}
\label{fig:prc_static_openai}
null
108
2508.00602
images/experiments/Static/static_toxicchat.pdf
\caption{ToxicChat dataset}
\label{fig:prc_static_toxic}
null
109
2508.00602
images/experiments/Static/static_pii.pdf
\caption{PII dataset}
\label{fig:prc_static_pii}
null
110
2508.00602
images/experiments/Static/static_openai.pdf
\caption{OpenAI dataset}
\label{fig:prc_static_openai}
null
111
2508.00602
images/experiments/Static/static_toxicchat.pdf
\caption{ToxicChat dataset}
\label{fig:prc_static_toxic}
null
112
2508.00602
images/experiments/Static/static_pii.pdf
\caption{PII dataset}
\label{fig:prc_static_pii}
null
113
2508.00602
images/appendices/cluster-example.png
\caption{Essential static LeakSealer Report.}
\label{fig:static-example-short}
null
114
2508.00602
images/experiments/Dynamic/dynamic_openai.pdf
\caption{OpenAI dataset}
\label{fig:prc_dynamic_openai}
null
115
2508.00602
images/experiments/Dynamic/dynamic_toxicchat.pdf
\caption{ToxicChat dataset}
\label{fig:prc_dynamic_toxic}
null
116
2508.00602
images/experiments/Dynamic/dynamic_pii.pdf
\caption{PII dataset}
\label{fig:prc_dynamic_pii}
null
117
2508.00602
images/experiments/Dynamic/dynamic_openai.pdf
\caption{OpenAI dataset}
\label{fig:prc_dynamic_openai}
null
118
2508.00602
images/experiments/Dynamic/dynamic_toxicchat.pdf
\caption{ToxicChat dataset}
\label{fig:prc_dynamic_toxic}
null
119
2508.00602
images/experiments/Dynamic/dynamic_pii.pdf
\caption{PII dataset}
\label{fig:prc_dynamic_pii}
null
120
2508.00602
images/appendices/cluster-example.png
\caption{Clusters from the ToxicChat test split. Static LeakSealer summary with outliers truncated to three examples.}
\label{fig:example-cluste}
null
121
2508.00228
figs/fig3-map.png
\caption{Network routes with different latencies interconnecting our DTNs}
\label{map}
null
122
2508.00228
figs/fig4-results-1-128-100.png
\caption{Network routes with different latencies interconnecting our DTNs}
\label{1-128-100}
null
123
2508.00228
figs/fig9-results-1-32-100.png
\caption{Distribution of throughput per number of streams across different RTTs}
\label{1-32-100}
null
124
2508.00228
figs/fig11-results-2-16-100.png
\caption{Total cores per server: 32, bandwidth limit: 100 Gbps}
\label{2-16-100}
null
125
2508.00228
figs/fig12-results-4-16-200.png
\caption{Total cores per server: 64, bandwidth limit: 200 Gbps}
\label{4-16-200}
null
126
2508.00228
figs/fig13-results-4-32-200.png
\caption{Total cores per server: 128, bandwidth limit: 200 Gbps}
\label{4-32-200}
null
127
2508.00228
figs/fig6-results-4-32-400.png
\caption{Total cores per server: 128, bandwidth limit: 400 Gbps}
\label{4-32-400}
null
128
2508.00228
figs/fig7-results-16-8-400.png
\caption{Network routes with different RTTs interconnecting our DTNs}
\label{16-8-400}
null
129
2508.00329
pics_s/LSA_a.jpg
\caption{\label{fig_LSA} Spatio-temporal profiles of perturbation $y_1$ showing that the mode with (a) $\alpha=2$ decays over time, while the mode with (b) $\alpha=10$ grows over time for $Q=80$ and $\delta=5$. }
null
null
130
2508.00329
pics_s/r_f.jpg
\caption{{\it The reaction front position evolution for different flow rate $Q$, diffusivity ratio $\delta$ and disturbance wavenumber $\alpha$. The green dashed lines indicate corresponding stationary front position $r_s$.}}
\label{rf_LSA}
null
131
2508.00329
pics_s/IL_dhminrel700_m.pdf
\caption{{\it (a) $E_{I,m}(t)$ and (b) $E_{V,m}(t)$ for different $m$ values. The rest of the parameters are fixed as $Q=80$, $\delta=10$.}}
\label{mgridtest}
null
132
2508.00329
pics_s/surfXYvsX_dtQ160Delta10t_.jpg
\caption{\label{surfXY_Qeffect} (a) Surface plots of the autocatalyst $X$ at different times for $Q=160,\delta=10$, (b) the spatio-temporal evolution of reactant $Y$ for the same parameter values.}
null
null
133
2508.00329
pics_s/temporal_num.jpg
\caption{ (a) Temporal evolution of the reaction front position, \( r_f \), for \( Q = 80 \) and \( Q = 160 \) when \( \delta = 10 \). (b) Temporal evolution of the wavelength, \( \lambda(t) \), for \( Q = 80 \) and \( Q = 160 \) when \( \delta = 10 \). Dashed lines indicate the mean wavelength, \( \lambda \). (c) Temporal evolution of the amplitude, \( A(t) \), for \( Q = 80 \) and \( Q = 160 \) when \( \delta = 10 \). Dashed lines indicate the mean amplitude, \( A \). }
\label{fig:temp_num}
null
134
2508.00329
pics_s/experimental_setup.pdf
\caption{\label{Fig: expset} Schematic of the experimental set-up.}
null
null
135
2508.00329
pics_s/figImageProcessing.pdf
\caption{\label{Fig: DataProcessing} (a) Original picture of the reactant solution Y, containing 30.6 mM of SPA, being continuously injected into the product solution X at a flow rate $\hat{Q}=1$ $\mu L$ $min^{-1}$, resulting in a stationary cellular front 10 mm away from the inlet. Scale bar: 10 mm. (b) Binarized image of the stationary front. (c) Outline of the front with the injection tube removed in the post-processing. (d) Comparison of the front position data in a radial coordinate set before and after FT-based filtering. The front's leading and trailing points are marked with upward and downward-facing triangles, respectively. The average front position is marked with a dashed red line.}
null
null
136
2508.00329
pics_s/temporal_exp.jpg
\caption{Temporal evolution of key observables in response to changes in flow rate $\hat{Q}$: (a) imposed flow rate $\hat{Q}$ over time, and resulting variations in (b) front position, (c) amplitude, and (d) wavelength during the experiments. Full blue lines represent experimental data, while dashed red lines represent their average value. While the dependence of the stationary front position $\hat{r}_s$ and the cellular modulation amplitude $\hat{A}$ on the flow rate $\hat{Q}$ are evident, there is no statistically significant relationship between the flow rate and the average modulation wavelength $\hat{\lambda}$.}
\label{fig:experiments_timelapse}
null
137
2508.05628v1
figs/segmentation_figure.pdf
\caption{Segmentation examples. H‑Net++ (bottom) correctly segments compound words and clitics, while BPE (top) over-fragments.}
\label{fig:segments}
null
138
2508.05628v1
figs/robustness_curves.pdf
\caption{Performance degradation under increasing ZWNJ corruption. H-Net++ maintains higher accuracy across all corruption levels.}
\label{fig:robustness}
null
139
2508.05628v1
figs/attention_patterns.pdf
\caption{Attention patterns in the Transformer mixer show long-range dependencies between morphologically related chunks.}
\label{fig:attention}
null
140
2508.00311
fig/ocr-fig-1.pdf
\caption{ Illustration of the main challenges in current LaTeX formula datasets: limited complexity and structural diversity. \textit{Left}: \textbf{Simple vs. Complex LaTeX Expressions.} Existing datasets primarily contain simple, single-structure formulas with few symbol types, which restricts model performance on advanced scientific content. In contrast, the proposed CSFormula dataset includes complex expressions with multi-level nesting, sophisticated subscripts and superscripts, and abundant scientific symbols, supporting more accurate recognition of diverse formulas. \textit{Right}: \textbf{Line-level vs. Page-level LaTeX Expressions.} Most datasets focus only on line-level formulas, overlooking the complex paragraph- and page-level layouts common in real documents. The proposed CSFormula dataset incorporates these higher-level structures, enabling models to better capture complex layouts and improving semantic reconstruction and document understanding. }
\label{fig:motivation}
null
141
2508.00311
fig/ocr-fig-2.pdf
\caption{Visualization of DocTron-Formula’s recognition results. The top shows accurate recognition of complex, deeply nested LaTeX expressions, while the bottom demonstrates robust parsing of page-level layouts.}
\label{fig:qualitative}
null
142
2508.00539
Cupritframework.png
\caption{Overall workflow of the proposed phase-locked SNR-based band selection and spectral unmixing pipeline. The process consists of three stages: (1) Phase-locked spectral smoothing and SNR-guided band selection; (2) unsupervised endmember extraction and NNLS-based unmixing; and (3) abundance map generation and spectral similarity assessment with reference mineral spectra.}
\label{fig:framework}
null
143
2508.00539
RGBmaps.png
\caption{RGB composite rendering of the Cuprite hyperspectral scene.}
\label{fig:hsi_rgb}
null
144
2508.00539
Labcupritespectral.png
\caption{Reference spectra for 12 minerals in the Cuprite district from USGS laboratory measurements.}
\label{fig:lab_spectra}
null
145
2508.00539
500pixels.png
\caption{Cuprite Spectral before Unmixing}
\label{fig:sg_comparison}
null
146
2508.00539
spectralsignaturecomparison.png
\caption{Comparison between Raw Cuprite Spectral and Auto-extracted Spectral from HSI.}
\label{fig:spectral_signature}
null
147
2508.00539
SNRSpectralBandRetainedBandsHighlighted.png
\caption{Band-wise SNR distribution with selected high-SNR bands highlighted.}
\label{fig:snr_profile}
null
148
2508.00539
abundancemaps.png
\caption{Abundance maps for Classes 0–2 using filtered input and NNLS.}
\label{fig:abundance_maps}
null
149
2508.00476
figures/amr_man_sell_flower.png
\caption{AMR graph for ``A man breaks a window.''}
\label{fig:amr_graph}
null
150
2508.00476
figures/auto_cz.png
\caption{Automatic evaluation using LLM-as-judges. Scores are between 0 and 10. Violin plot with mean distribution as blue line. *** for p$\leq$0.005}
\label{fig:auto_cz}
null
151
2508.00476
figures/auto_en.png
\caption{Automatic evaluation using LLM-as-judges. Scores are between 2 and 10. Violin plot with distribution mean as blue line. *** for p$\leq$0.005}
\label{fig:auto_en}
null
152
2508.00476
figures/manual_en.png
\caption{Human evaluation. Scores are between 2 and 10. Violin plot with distribution mean as blue line. *** for p$\leq$0.005}
\label{fig:manual_en}
null
295
2508.00667
logc_est_bias_2.png
\caption{Same as Figure~\ref{fig:logM_est_bias}, but shows the lensing estimator bias of $\log \hat{c}$. }
\label{fig:logc_est_bias}
null
153
2508.00476
figures/auto_vs_manual_en.png
\caption{Human versus automatic evaluation. Scores are between 2 and 10. Violin plot with distribution mean as blue line. *** for p$\leq$0.005}
\label{fig:manual_auto_en}
null
154
2508.00721
fig/chi2LL.png
\caption{Plot of the function $h(\mb z_0)$ (after a change of variable $u = \norm{\mb z_0}_2^2$). An ideal regularization function should blow up sharply away from the narrow concentration region in orange to promote Gaussianity effectively.}
\label{fig:chi2LL}
null
155
2508.00721
fig/visualization/sr-afhq-measurement.png
\caption{Measurement}
null
null
156
2508.00721
fig/visualization/sr-afhq-gt.png
\caption{Ground Truth}
null
null
157
2508.00721
fig/visualization/sr-afhq-dip.png
\caption{DIP}
null
null
158
2508.00721
fig/visualization/sr-afhq-flowchef.png
\caption{FlowChef}
null
null
159
2508.00721
fig/visualization/sr-afhq-flowdps.png
\caption{FlowDPS}
null
null
160
2508.00721
fig/visualization/sr-afhq-dflow.png
\caption{D-Flow}
null
null
161
2508.00721
fig/visualization/sr-afhq-fmplug-w.png
\caption{FMPlug-W}
null
null
162
2508.00721
fig/visualization/sr-afhq-fmplug-w-r.png
\caption{FMPlug-W-R}
null
null
163
2508.00721
fig/visualization/sr-afhq-measurement.png
\caption{Measurement}
null
null
164
2508.00721
fig/visualization/sr-afhq-gt.png
\caption{Ground Truth}
null
null
165
2508.00721
fig/visualization/sr-afhq-dip.png
\caption{DIP}
null
null
166
2508.00721
fig/visualization/sr-afhq-flowchef.png
\caption{FlowChef}
null
null
167
2508.00721
fig/visualization/sr-afhq-flowdps.png
\caption{FlowDPS}
null
null
168
2508.00721
fig/visualization/sr-afhq-dflow.png
\caption{D-Flow}
null
null
169
2508.00721
fig/visualization/sr-afhq-fmplug-w.png
\caption{FMPlug-W}
null
null
170
2508.00721
fig/visualization/sr-afhq-fmplug-w-r.png
\caption{FMPlug-W-R}
null
null
171
2508.00721
fig/visualization/sr-div-measurement.png
\caption{Measurement}
null
null
172
2508.00721
fig/visualization/sr-div-gt.png
\caption{Ground Truth}
null
null
173
2508.00721
fig/visualization/sr-div-dip.png
\caption{DIP}
null
null
174
2508.00721
fig/visualization/sr-div-flowchef.png
\caption{FlowChef}
null
null
175
2508.00721
fig/visualization/sr-div-flowdps.png
\caption{FlowDPS}
null
null
176
2508.00721
fig/visualization/sr-div-dflow.png
\caption{D-Flow}
null
null
177
2508.00721
fig/visualization/sr-div-fmplug-w.png
\caption{FMPlug-W}
null
null
178
2508.00721
fig/visualization/sr-div-fmplug-w-r.png
\caption{FMPlug-W-R}
null
null
179
2508.00721
fig/visualization/sr-div-measurement.png
\caption{Measurement}
null
null
180
2508.00721
fig/visualization/sr-div-gt.png
\caption{Ground Truth}
null
null
181
2508.00721
fig/visualization/sr-div-dip.png
\caption{DIP}
null
null
182
2508.00721
fig/visualization/sr-div-flowchef.png
\caption{FlowChef}
null
null
183
2508.00721
fig/visualization/sr-div-flowdps.png
\caption{FlowDPS}
null
null
184
2508.00721
fig/visualization/sr-div-dflow.png
\caption{D-Flow}
null
null
185
2508.00721
fig/visualization/sr-div-fmplug-w.png
\caption{FMPlug-W}
null
null
186
2508.00721
fig/visualization/sr-div-fmplug-w-r.png
\caption{FMPlug-W-R}
null
null
187
2508.00721
fig/visualization/gd-afhq-measurement.png
\caption{Measurement}
null
null
188
2508.00721
fig/visualization/gd-afhq-gt.png
\caption{Ground Truth}
null
null
189
2508.00721
fig/visualization/gd-afhq-dip.png
\caption{DIP}
null
null
190
2508.00721
fig/visualization/gd-afhq-flowchef.png
\caption{FlowChef}
null
null
191
2508.00721
fig/visualization/gd-afhq-flowdps.png
\caption{FlowDPS}
null
null
192
2508.00721
fig/visualization/gd-afhq-dflow.png
\caption{D-Flow}
null
null
193
2508.00721
fig/visualization/gd-afhq-fmplug-w.png
\caption{FMPlug-W}
null
null
194
2508.00721
fig/visualization/gd-afhq-fmplug-w-r.png
\caption{FMPlug-W-R}
null
null
195
2508.00721
fig/visualization/gd-afhq-measurement.png
\caption{Measurement}
null
null
196
2508.00721
fig/visualization/gd-afhq-gt.png
\caption{Ground Truth}
null
null
197
2508.00721
fig/visualization/gd-afhq-dip.png
\caption{DIP}
null
null
198
2508.00721
fig/visualization/gd-afhq-flowchef.png
\caption{FlowChef}
null
null