Add paper link, pipeline tag and improve model card

Hi, I'm Niels from the Hugging Face community science team. I'm opening this PR to improve the model card for MATT-Diff.

This PR adds:
- The `robotics` pipeline tag to the metadata.
- A link to the [MATT-Diff paper](https://huggingface.co/papers/2511.11931).
- Links to the official [GitHub repository](https://github.com/CINAPSLab/MATT-Diff).
- Detailed usage instructions for downloading and evaluating the pre-trained checkpoints.
- The BibTeX citation for the paper.

README.md

@@ -1,4 +1,47 @@
 ---
 license: apache-2.0
+pipeline_tag: robotics
 ---
-Models for MATT-Diff(https://github.com/CINAPSLab/MATT-Diff)
+
+# MATT-Diff: Multimodal Active Target Tracking by Diffusion Policy
+
+This repository contains the models presented in the paper [MATT-Diff: Multimodal Active Target Tracking by Diffusion Policy](https://huggingface.co/papers/2511.11931).
+
+MATT-Diff is a diffusion-policy-based framework for **active multi-target tracking** in partially observable indoor environments. It learns multi-modal action strategies from three expert planners, using a transformer map encoder and an attention-based target encoder over Gaussian beliefs.
+
+- **Paper:** [https://huggingface.co/papers/2511.11931](https://huggingface.co/papers/2511.11931)
+- **Repository:** [https://github.com/CINAPSLab/MATT-Diff](https://github.com/CINAPSLab/MATT-Diff)
+
+## Overview
+The policy enables multiple behavior modes for the agent, including exploration, tracking, and target reacquisition, without prior knowledge of the target numbers, states, or dynamics. Trained as a diffusion model, the policy learns to generate multimodal action sequences through a denoising process.
+
+## Usage
+
+### Download Checkpoints
+You can download the pretrained checkpoints using the Hugging Face CLI:
+
+```bash
+huggingface-cli download saidaliu27/MATT-Diff dp/best.pt --local-dir src/run/dp
+huggingface-cli download saidaliu27/MATT-Diff bc/best.pt --local-dir src/run/bc
+```
+
+### Evaluate
+After downloading the checkpoints and setting up the environment as described in the [GitHub repository](https://github.com/CINAPSLab/MATT-Diff), you can evaluate MATT-Diff:
+
+```bash
+python -u -m evals.eval_dp
+```
+
+## Citation
+
+```bibtex
+@misc{liu2025mattdiffmultimodalactivetarget,
+      title={MATT-Diff: Multimodal Active Target Tracking by Diffusion Policy}, 
+      author={Saida Liu and Nikolay Atanasov and Shumon Koga},
+      year={2025},
+      eprint={2511.11931},
+      archivePrefix={arXiv},
+      primaryClass={cs.RO},
+      url={https://arxiv.org/abs/2511.11931}, 
+}
+```