DyCo-RL: Dynamic Cross-Modal Coordination for Visual Reasoning

• URL: http://arxiv.org/abs/2606.08035v1
• Date: Sat, 06 Jun 2026 07:56:25 GMT
• Title: DyCo-RL: Dynamic Cross-Modal Coordination for Visual Reasoning
• Abstract summary: We propose DyCo-RL, which integrates dynamic cross-modal coordination into RLVR optimization.<n>Experiments demonstrate that DyCo-RL, when applied to Qwen2.5-VL-3B/7B, consistently improves four representative RLVR algorithms.
• Score: 49.985643703496514
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a leading paradigm for enhancing visual reasoning in Multimodal Large Language Models (MLLMs). However, existing RLVR methods optimize primarily for the reasoning outcome, fundamentally overlooking the fine-grained cross-modal coordination required during the generation process. Through token-level analyses and controlled interventions, we reveal that during Chain-of-Thought (CoT) reasoning, MLLMs frequently fail to dynamically alternate between extracting visual evidence and synthesizing textual context-a coordination breakdown that is causally linked to reasoning failures. Motivated by these findings, we propose DyCo-RL, which integrates dynamic cross-modal coordination into RLVR optimization. Specifically, DyCo-RL uses the Fisher-Rao geodesic distance to measure within-modality attention shifts, assigning tokens to either visually-oriented or text-oriented functional roles. It then evaluates the alignment between a token's actual attention allocation and its assigned role, leveraging this score for alignment-guided advantage reweighting during policy optimization. Extensive experiments demonstrate that the algorithm-agnostic DyCo-RL, when applied to Qwen2.5-VL-3B/7B, consistently improves four representative RLVR algorithms across seven benchmarks spanning visual-centric and mathematical reasoning.

Related papers

• CORA: Analyzing and bridging thinking-answer gap in Multimodal RLVR via Consistency-Oriented Reasoning Alignment [32.29376299423774]
  Reinforcement learning with verifiable rewards (RLVR) has successfully elicited the reasoning capabilities of large language models.<n>Existing methods primarily focus on improving the visual coverage of reasoning traces and mitigating visual hallucinations, but underestimate the semantic inconsistency between the reasoning process and the final answer.<n>We propose Consistency-Oriented Reasoning Alignment (CORA), which introduces thinking-answer semantic consistency into RLVR through a lightweight plug-and-play consistency reward model.
  arXiv  Detail & Related papers  (2026-06-12T17:54:59Z)
• CurveRL: Principled Distribution-Aware Context Reweighting for LLM Reasoning [20.44110921033217]
  Context or prompt-level reweighting has emerged as a central algorithmic lever in Reinforcement Learning with Verified Rewards.<n>We propose a distribution-aware prompt reweighting approach, called CurveRL, based on a quantile coordinate transform.<n>Our study identifies context-distribution control as a principled axis for analyzing and designing prompt-reweighted RLVR algorithms.
  arXiv  Detail & Related papers  (2026-05-23T01:18:55Z)
• Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMs [51.60575965819268]
  We propose a plug-and-play Token-Reweighting (ToR) strategy that explicitly models this interdependence.<n>ToR delivers consistent performance gains across multiple multi-modal reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-26T06:25:27Z)
• Rethinking Token-Level Policy Optimization for Multimodal Chain-of-Thought [73.39221516441624]
  Multimodal Chain-of-Thought (CoT) reasoning requires large vision-language models to construct reasoning trajectories.<n>Existing Reinforcement Learning with Verifiable Rewards (RLVR) methods treat CoT uniformly without distinguishing varying degrees of visual grounding.<n>We propose Perception-Exploration Policy Optimization (PEPO), which derives a perception prior to hidden state similarity and integrates it with token entropy.
  arXiv  Detail & Related papers  (2026-03-24T06:38:00Z)
• Contextual Rollout Bandits for Reinforcement Learning with Verifiable Rewards [69.74686029941881]
  Reinforcement Learning with Verifiable Rewards (RLVR) is an effective paradigm for improving the reasoning capabilities of large language models.<n>We propose a unified neural scheduling framework that adaptively selects high-value rollouts throughout training.<n>Experiments on six mathematical reasoning benchmarks demonstrate consistent gains in performance and training efficiency across multiple RLVR optimization methods.
  arXiv  Detail & Related papers  (2026-02-09T10:51:58Z)
• Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration [49.9937230730202]
  We propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention.<n>Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories.<n>We show that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales.
  arXiv  Detail & Related papers  (2026-02-03T15:32:09Z)
• Spotlight on Token Perception for Multimodal Reinforcement Learning [65.97597482517425]
  Reinforcement Learning with Verifiable Rewards (RLVR) has advanced the reasoning capabilities of Large Vision-Language Models (LVLMs)<n>In this paper, we undertake a pioneering exploration of multimodal RLVR through the novel perspective of token perception.<n>We propose Visually-Perceptive Policy Optimization (VPPO), a novel policy gradient algorithm that explicitly leverages token perception to refine the learning signal.
  arXiv  Detail & Related papers  (2025-10-10T11:25:33Z)
• Cross-Trajectory Representation Learning for Zero-Shot Generalization in RL [21.550201956884532]
  generalize policies learned on a few tasks over a high-dimensional observation space to similar tasks not seen during training. Many promising approaches to this challenge consider RL as a process of training two functions simultaneously. We propose Cross-Trajectory Representation Learning (CTRL), a method that runs within an RL agent and conditions its encoder to recognize behavioral similarity in observations.
  arXiv  Detail & Related papers  (2021-06-04T00:43:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.