Rethinking Token-Level Policy Optimization for Multimodal Chain-of-Thought

• URL: http://arxiv.org/abs/2603.22847v1
• Date: Tue, 24 Mar 2026 06:38:00 GMT
• Title: Rethinking Token-Level Policy Optimization for Multimodal Chain-of-Thought
• Abstract summary: 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.
• Score: 73.39221516441624
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Multimodal Chain-of-Thought (CoT) reasoning requires large vision-language models to construct reasoning trajectories that interleave perceptual grounding with multi-step inference. However, existing Reinforcement Learning with Verifiable Rewards (RLVR) methods typically optimize reasoning at a coarse granularity, treating CoT uniformly without distinguishing their varying degrees of visual grounding. In this work, we conduct a token-level analysis of multimodal reasoning trajectories and show that successful reasoning is characterized by structured token dynamics reflecting both perceptual grounding and exploratory inference. Building upon this analysis, we propose Perception-Exploration Policy Optimization (PEPO), which derives a perception prior from hidden state similarity and integrates it with token entropy through a smooth gating mechanism to produce token-level advantages. PEPO integrates seamlessly with existing RLVR frameworks such as GRPO and DAPO, requiring neither additional supervision nor auxiliary branches. Extensive experiments across diverse multimodal benchmarks demonstrate consistent and robust improvements over strong RL baselines, spanning geometry reasoning, visual grounding, visual puzzle solving, and few-shot classification, while maintaining stable training dynamics. Code: https://github.com/xzxxntxdy/PEPO

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)
• Enhancing Multimodal In-Context Learning via Inductive-Deductive Reasoning [65.15766304205657]
  In-context learning (ICL) allows large models to adapt to tasks using a few examples, yet its extension to vision-language models (VLMs) remains fragile.<n>Our analysis reveals that the fundamental limitation lies in an inductive gap, models often produce correct answers from flawed reasoning.<n>We introduce a framework that restructures multimodal ICL as a principled inductive-deductive process.
  arXiv  Detail & Related papers  (2026-05-04T09:18:19Z)
• Not All Tokens See Equally: Perception-Grounded Policy Optimization for Large Vision-Language Models [38.47027398567909]
  Perception-Grounded Policy Optimization (PGPO) is a novel fine-grained credit assignment framework that dynamically reshapes advantages at the token level.<n>We show that PGPO actively amplifies learning signals for visually-dependent tokens while suppressing gradient noise from linguistic priors.<n>Both theoretical and empirical analyses confirm that PGPO effectively reduces gradient variance, prevents training collapse, and acts as a potent regularizer for robust, perception-grounded multimodal reasoning.
  arXiv  Detail & Related papers  (2026-04-02T09:53:20Z)
• 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)
• Perception-Aware Multimodal Spatial Reasoning from Monocular Images [57.42071289037214]
  spatial reasoning from monocular images is essential for autonomous driving.<n>Current Vision-Language Models (VLMs) still struggle with fine-grained geometric perception.<n>We propose a perception-aware multimodal reasoning framework that equipsVLMs with explicit object-centric grounding ability.
  arXiv  Detail & Related papers  (2026-03-07T02:05:12Z)
• CASHEW: Stabilizing Multimodal Reasoning via Iterative Trajectory Aggregation [6.356820150960838]
  We introduce two complementary approaches inspired by test-time scaling to stabilize vision-language models.<n>CASHEW is an inference-time framework that stabilizes reasoning by iteratively aggregating multiple candidate trajectories into higher-quality reasoning traces.<n>CASHEW-RL is trained using Group Sequence Policy Optimization (GSPO) with a composite reward that encourages correct answers grounded in minimal yet sufficient visual evidence.
  arXiv  Detail & Related papers  (2026-01-12T21:24:45Z)
• Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space [46.05748768260013]
  We propose a test-time Dynamic Multimodal Latent Reasoning framework.<n>It employs confidence-guided latent policy gradient optimization to latent think tokens for in-depth reasoning.<n> Experiments across seven multimodal reasoning benchmarks and various model architectures demonstrate that DMLR significantly improves reasoning and perception performance.
  arXiv  Detail & Related papers  (2025-12-14T10:07:45Z)
• Multi-Path Collaborative Reasoning via Reinforcement Learning [54.8518809800168]
  Chain-of-Thought (CoT) reasoning has significantly advanced the problem-solving capabilities of Large Language Models (LLMs)<n>Recent methods attempt to address this by generating soft abstract tokens to enable reasoning in a continuous semantic space.<n>We propose Multi-Path Perception Policy Optimization (M3PO), a novel reinforcement learning framework that explicitly injects collective insights into the reasoning process.
  arXiv  Detail & Related papers  (2025-12-01T10:05:46Z)
• 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)

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.