README.md

---
license: apache-2.0
base_model:
- Qwen/Qwen2.5-7B-Instruct
tags:
  - vlm
  - chart-understanding
library_name: transformers
---

# BiPS — Bi-directional Perceptual Shaping for Multimodal Reasoning

This model card describes **BiPS (Bi-directional Perceptual Shaping)**, a **training-time** framework proposed in *“See Less, See Right: Bi-directional Perceptual Shaping For Multimodal Reasoning”* **[CVPR 2026]**.

- Paper: https://arxiv.org/abs/2512.22120  
- Code: https://github.com/zss02/BiPS

## What is BiPS?

Many VLMs fail on multimodal reasoning because they **look at the wrong visual evidence** (especially for charts, thin lines, intersections, and small regions).
BiPS transforms visual cues into training-time policy constraints by constructing evidence-preserving and evidence-ablated views, enabling the model to internalize fine-grained perception without inference-time overhead. This formulation can also be interpreted from the perspective of visual on-policy distillation (OPD) / on-policy self-distillation, where information-asymmetric visual views provide dense policy-shaping signals for GRPO training.

## Key idea

BiPS trains a VLM with two complementary view transformations:

- **Evidence-Preserving View**: keep only the visual evidence needed to answer, reduce distractions.  
  → enforce **consistency** between predictions from the original image and the preserved view.

- **Evidence-Ablated View**: remove the key evidence so the image no longer supports the answer.  
  → enforce **separation** so the model cannot rely on shortcuts.

These constraints are typically implemented with **KL-based objectives** and can be integrated into **GRPO** training.

## Why it matters

- Better **fine-grained evidence alignment**
- Less “guessing” from language priors
- **No additional inference overhead** (views are used only during training)

## How to use

BiPS is mainly a **training recipe**. To apply it:
1. Follow the official repo to set up dependencies and scripts.
2. Train your base VLM with BiPS-generated **preserve/ablate** views.
3. Use the resulting checkpoint as a standard VLM at inference time (no extra steps).

## Citation

```bibtex
@article{zhang2025bips,
  title={See Less, See Right: Bi-directional Perceptual Shaping For Multimodal Reasoning},
  author={Zhang, Shuoshuo and Zhang, Yizhen and Fu, Jingjing and Song, Lei and Bian, Jiang and Yang, Yujiu and Wang, Rui},
  journal={arXiv preprint arXiv:2512.22120},
  year={2025}
}
```