Update README.md

README.md

@@ -23,4 +23,17 @@ size_categories:
 We adopt the official split:  
 - Train:  1,200 videos, 48,774 captions  
 - Val: 100 videos, 4,290 captions  
-- Test: 670 videos, 27,763 captions  
+- Test: 670 videos, 27,763 captions
+
+---
+
+## 🌟 Citation
+
+```bibtex
+@inproceedings{chen2011collecting,
+  title={Collecting highly parallel data for paraphrase evaluation},
+  author={Chen, David and Dolan, William B},
+  booktitle={Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL)},
+  year={2011}
+}
+```