Does Seeing More Mean Knowing More? Mono-Anchored Advantage Normalization for Multi-Source Visual Reasoning
2026-05-25 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address a problem with using multiple types of visual data at once, like infrared and depth images, where combining them can sometimes create confusion rather than help. They propose a new method called MARS that treats each type of visual data separately and uses single-source results as a reference point to check if adding more data actually helps. This approach helps the system focus on useful information and reduce noise when learning through reinforcement learning techniques. Their tests show that MARS improves performance compared to older methods on different datasets.
reinforcement learningvisual reasoningmulti-source fusioninformation gainadvantage normalizationmodalitygradient estimationinfrared imagingdepth imagingmulti-modal learning
Authors
Fanhu Zeng, Zhicong Luo, Zefan Wang, You Li, Chi Chen, Maosong Sun
Abstract
Visual reasoning through reinforcement learning with verifiable rewards (RLVR) has achieved remarkable progress. However, when dealing with multi-source inputs, existing approaches tend to treat them as a mere accumulation of information, lacking explicit mechanisms to distinguish whether integrating additional sources yields information gain or introduces interference. Therefore, they struggle to effectively model dynamic interaction when integrating multiple sources, particularly when they differ significantly in physical properties and semantics, e.g., infrared and depth, leading to inferior performance to mono-source reasoning when a certain source holds the dominant signal. To address this issue, we propose MARS, a novel mono-anchored multi-source reasoning framework that models each visual modality as an independent information source. Specifically, by treating mono-source rewards as dynamic anchors, our method explicitly incorporates the information gain introduced by multi-source fusion into advantage normalization and adaptively emphasizes mutual promotion between sources while suppressing potential noise or conflicts during RLVR. From theoretical analysis, our method effectively quantifies information gain introduced by multi-source integration in gradient estimation, enabling consistent modality regulation. Empirical results also show impressive 3.2% and 4.9% performance gains on GRPO and DAPO across diverse datasets, confirming effectiveness of our method.