Improve dataset card: Add task categories, license, links, and description

This PR enhances the dataset card for `physical-ai-bench-understanding` by:
- Adding `task_categories: video-text-to-text` to improve discoverability.
- Specifying the `license: mit` as indicated in the GitHub repository.
- Including a brief introduction to the PAI-Bench project and its purpose.
- Providing direct links to the associated paper ([PAI-Bench: A Comprehensive Benchmark For Physical AI](https://huggingface.co/papers/2512.01989)) and the GitHub repository ([SHI-Labs/physical-ai-bench](https://github.com/SHI-Labs/physical-ai-bench)).

These changes will make the dataset card more informative and align it with best practices for dataset documentation on the Hugging Face Hub.

README.md

@@ -1,4 +1,7 @@
 ---
+task_categories:
+- video-text-to-text
+license: mit
 dataset_info:
   features:
   - name: question
@@ -36,6 +39,11 @@ configs:
 
 # Physical AI Bench - Understanding
 
+PAI-Bench (Physical AI Bench) is a comprehensive benchmark designed to evaluate physical AI generation and understanding capabilities across various real-world scenarios. This particular dataset, **PAI-Bench-U**, focuses specifically on **Video Understanding** tasks, comprising 2,808 real-world cases with task-aligned metrics.
+
+- **Paper:** [PAI-Bench: A Comprehensive Benchmark For Physical AI](https://huggingface.co/papers/2512.01989)
+- **Code:** [GitHub Repository](https://github.com/SHI-Labs/physical-ai-bench)
+
 ## Citation
 
 If you use Physical AI Bench in your research, please cite:
@@ -43,11 +51,11 @@ If you use Physical AI Bench in your research, please cite:
 ```bibtex
 @misc{zhou2025paibenchcomprehensivebenchmarkphysical,
       title={PAI-Bench: A Comprehensive Benchmark For Physical AI}, 
-      author={Fengzhe Zhou and Jiannan Huang and Jialuo Li and Deva Ramanan and Humphrey Shi},
-      year={2025},
-      eprint={2512.01989},
-      archivePrefix={arXiv},
-      primaryClass={cs.CV},
-      url={https://arxiv.org/abs/2512.01989}, 
+      author={Fengzhe Zhou and Jiannan Huang and Jialuo Li and Deva Ramanan and Humphrey Shi},\
+      year={2025},\
+      eprint={2512.01989},\
+      archivePrefix={arXiv},\
+      primaryClass={cs.CV},\
+      url={https://arxiv.org/abs/2512.01989}, \
 }
-```
+```