Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

IVR-R1: Refining Trajectories through Iterative Visual-Grounded Reasoning in Reinforcement Learning

About

Multimodal large language models via reinforcement learning (RL) have demonstrated remarkable capabilities in complex visual reasoning tasks, yet they remain limited in long-horizon multimodal scenarios, often suffering from visual hallucination and logical error. Current methods typically pre-encode high-dimensional visual scenes into discrete textual proxies to facilitate downstream reasoning. As the reasoning chain unfolds, however, the inherent information asymmetry between text and visual scenes tends to erode visual grounding, resulting in misguided reasoning and erroneous outputs. To address this issue, we introduce IVR-R1 (Iterative Visual-grounded Reasoning), a novel RL training framework that facilitates dynamic visual re-alignment that actively rectifies reasoning trajectories to guide policy optimization. Specifically, by leveraging a reward-driven screening mechanism to identify flawed rollouts, IVR-R1 executes a fine-grained, step-level error attribution within the multimodal context. By iteratively cross-referencing intermediate reasoning states against pristine visual priors, a Re-Reasoning Loop enables automated trajectory rectification, effectively synthesizing expert-level demonstrations that serve as high-fidelity reasoning templates for the policy model. Our experiments across diverse multimodal benchmarks demonstrate that IVR-R1 consistently outperforms existing reinforcement learning methods, establishing a superior paradigm for maintaining logical and visual consistency in complex multimodal reasoning.

Chenghao Li, Fusheng Hao, Xikai Zhang, Likang Xiao, Yanwei Ren, Fuxiang Wu, Quan Chen, Liu Liu• 2026

Related benchmarks

TaskDatasetResultRank
Mathematical ReasoningMathVerse
Accuracy49.4
183
Multimodal UnderstandingMMMU
Accuracy (MMMU)56
52
Mathematical ReasoningMathVision mini (test)
Accuracy35.2
44
Mathematical ReasoningMathVista
Accuracy (%)69.6
29
General Multimodal UnderstandingMMMU-Pro
Accuracy53.3
13
General-purpose Multimodal UnderstandingMMStar
Accuracy63.3
13
Showing 6 of 6 rows

Other info

Follow for update