ChainV: Atomic Visual Hints Make Multimodal Reasoning Shorter and Better
This work addresses efficiency and accuracy issues in multimodal reasoning for tasks like math-intensive benchmarks, representing an incremental improvement over existing methods.
The paper tackles the problem of redundant self-reflection in multimodal reasoning models by proposing ChainV, a framework that dynamically integrates visual hints to shorten reasoning chains, resulting in a 2.3% accuracy improvement on MathVista, 51.4% reduction in inference latency, and 24.5% shorter output tokens.
Recent advances in multimodal reasoning models have demonstrated impressive capabilities across text and vision. However, even leading models exhibit redundant self-reflection when generating lengthy reasoning chains. While training-free CoT compression methods have emerged in the LLMs domain, they rely on static visual references and thus provide limited gains for multimodal reasoning. Therefore, we propose ChainV, a framework that dynamically integrates visual hints into the reasoning process, thereby making multimodal reasoning shorter and better. Specifically, ChainV first performs a coarse visual patch selection based on the previous reasoning step, then refines it by identifying the most representative atomic visual hint according to the averaged attention intensity. Additionally, ChainV introduces a consistency-based evaluation mechanism to assess the reliability of the chosen hint, guiding the model to adaptively adjust its level of self-reflection. Eventually, the pixel coordinates of the selected visual hint and its reliability are incorporated into thinking with a Bernoulli stochastic process. Experiments indicate that our method significantly improves reasoning accuracy and efficiency, especially on math-intensive benchmarks where visual hints are crucial for multi-step symbolic reasoning. For example, ChainV achieves $2.3\%$ improvement on the MathVista within MIMO-VL-RL, while reducing inference latency by $51.4\%$ and shortening output token length by $24.5\%$.