SRPO: Enhancing Multimodal LLM Reasoning via Reflection-Aware Reinforcement Learning

TL;DR

This paper introduces SRPO, a reinforcement learning framework that improves multimodal large language models' reasoning by incorporating explicit self-reflection and self-correction, leading to better accuracy and reflection quality.

Contribution

The paper proposes a novel two-stage reflection-aware RL framework, SRPO, which enhances multimodal LLM reasoning through high-quality reflection data and a new reward mechanism.

Findings

SRPO outperforms state-of-the-art models on multiple benchmarks.

Significant improvements in reasoning accuracy.

Enhanced reflection quality and self-correction capabilities.

Abstract

Multimodal large language models (MLLMs) have shown promising capabilities in reasoning tasks, yet still struggle with complex problems requiring explicit self-reflection and self-correction, especially compared to their unimodal text-based counterparts. Existing reflection methods are simplistic and struggle to generate meaningful and instructive feedback, as the reasoning ability and knowledge limits of pre-trained models are largely fixed during initial training. To overcome these challenges, we propose Multimodal Self-Reflection enhanced reasoning with Group Relative Policy Optimization (SRPO), a two-stage reflection-aware reinforcement learning (RL) framework explicitly designed to enhance multimodal LLM reasoning. In the first stage, we construct a high-quality, reflection-focused dataset under the guidance of an advanced MLLM, which generates reflections based on initial responses to help the policy model learn both reasoning and self-reflection. In the second stage, we introduce a novel reward mechanism within the GRPO framework that encourages concise and cognitively meaningful reflection while avoiding redundancy. Extensive experiments across multiple multimodal reasoning benchmarks, including MathVista, MathVision, MathVerse, and MMMU-Pro, using Qwen-2.5-VL-7B and Qwen-2.5-VL-32B demonstrate that SRPO significantly outperforms state-of-the-art models, achieving notable improvements in both reasoning accuracy and reflection quality.