paper-with-me

홈 › Papers

Stable and Efficient Single-Rollout RL for Multimodal Reasoning

2025-12-20 · Rui Liu, Dian Yu, Lei Ke, Haolin Liu, Yujun Zhou, Zhenwen Liang, Haitao Mi, Pratap Tokekar, Dong Yu arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) has become a key paradigm to improve the reasoning capabilities of Multimodal Large Language Models (MLLMs). However, prevalent group-based algorithms such as GRPO require multi-rollout sampling for each prompt. While more efficient single-rollout variants have recently been explored in text-only settings, we find that they suffer from severe instability in multimodal contexts, often leading to training collapse. To address this training efficiency-stability trade-off, we introduce $\textbf{MSSR}$ (Multimodal Stabilized Single-Rollout), a group-free RLVR framework that achieves both stable optimization and effective multimodal reasoning performance. MSSR achieves this via an entropy-based advantage-shaping mechanism that adaptively regularizes advantage magnitudes, preventing collapse and maintaining training stability. While such mechanisms have been used in group-based RLVR, we show that in the multimodal single-rollout setting they are not merely beneficial but essential for stability. In in-distribution evaluations, MSSR demonstrates superior training compute efficiency, achieving similar validation accuracy to the group-based baseline with half the training steps. When trained for the same number of steps, MSSR's performance surpasses the group-based baseline and shows consistent generalization improvements across five diverse reasoning-intensive benchmarks. Together, these results demonstrate that MSSR enables stable, compute-efficient, and effective RLVR for complex multimodal reasoning tasks.

📄 PDF Abstract BibTeX arXiv:2512.18215

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement LearningMultimodal Reasoning

Similar Papers 제목 키워드 기반

Consolidating Reinforcement Learning for Multimodal Discrete Diffusion Models

2025-10-03 · Tianren Ma, Mu Zhang, Yibing Wang, Qixiang Ye arxiv

Optimizing discrete diffusion model (DDM) with rewards remains a challenge: the non-autoregressive paradigm makes importance sampling intractable and rollout complex, puzzling reinforcement learning methods such as Group…

Reinforcement Learning

Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States

2026-05-08 · Yunho Choi, Jongwon Lim, Woojin Ahn, Minjae Oh 외 arxiv

Reinforcement learning with verifiable rewards (RLVR) for Large Reasoning Models hinges on baseline estimation for variance reduction, but existing approaches pay a heavy price: PPO requires a policy-model scale critic, …

Reinforcement Learning

DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning

2026-06-13 · David Huang, Lianlei Shan arxiv

Reasoning capabilities of multimodal large language models (MLLMs) have improved considerably in recent years. Existing approaches typically rely on explicit chain-of-thought or continuous latent-space trajectories to en…

Reinforcement LearningMultimodal Reasoning

PLUME: Latent Reasoning Based Universal Multimodal Embedding

2026-04-02 · Chenwei He, Xiangzhao Hao, Tianyu Yang, Yuxiang Ma 외 arxiv

Universal multimodal embedding (UME) maps heterogeneous inputs into a shared retrieval space with a single model. Recent approaches improve UME by generating explicit chain-of-thought (CoT) rationales before extracting e…

Switch-Reasoner: Learn When to Think in Multitask Mixtures via Reinforcement Learning

2026-07-09 · Yiyang Fang, Pei Fu, Jinjie Li, Jian Liang 외 arxiv

Multimodal Large Language Models (MLLMs) often follow a fixed Think-then-Answer paradigm, which is inefficient in heterogeneous multitask settings because simple inputs may not require explicit reasoning while difficult …

Reinforcement Learning