---
license: apache-2.0
task_categories:
- image-text-to-text
tags:
- visual-reasoning
- reinforcement-learning
- multi-source
- rlvr
---

# MARS: Mono-Anchored Multi-Source Visual Reasoning

This repository contains the dataset for the paper "[Does Seeing More Mean Knowing More? Mono-Anchored Advantage Normalization for Multi-Source Visual Reasoning](https://huggingface.co/papers/2605.25437)".

- **GitHub Repository:** [https://github.com/AI9Stars/MARS](https://github.com/AI9Stars/MARS)
- **Paper:** [https://arxiv.org/abs/2605.25437](https://arxiv.org/abs/2605.25437)

## Introduction
MARS is a novel mono-anchored multi-source reasoning framework designed for Reinforcement Learning with Verifiable Rewards (RLVR). It treats each visual modality as an independent information source and uses mono-source rewards as dynamic anchors to explicitly incorporate information gain from multi-source fusion while suppressing potential noise or conflicts.

## Dataset Summary
The dataset provides reinforcement fine-tuning (RL) data for several multi-source visual reasoning domains:

| Domain | Task Type |
| :--- | :--- |
| **Infrared** | Grounding tasks (RGB & IR) |
| **Depth** | Spatial reasoning using depth information |
| **Multi-view** | Reasoning across multiple perspectives |
| **Text-rich** | Visual reasoning in document-intensive environments |

For instructions on how to download the original images and prepare the data for training using the MARS framework, please refer to the [official GitHub repository](https://github.com/AI9Stars/MARS).

## Citation
If you find this work useful, please cite our paper:

```bibtex
@article{zeng2026does,
  title={Does Seeing More Mean Knowing More? Mono-Anchored Advantage Normalization for Multi-Source Visual Reasoning},
  author={Zeng, Fanhu and Luo, Zhicong and Wang, Zefan and Li, You and Chen, Chi and Sun, Maosong},
  journal={arXiv preprint arXiv:2605.25437},
  year={2026}
}
```