Guided Verifier: Collaborative Multimodal Reasoning via Dynamic Process Supervision

TL;DR

This paper introduces the Guided Verifier framework that enhances multimodal reasoning in large language models by integrating real-time dynamic verification, reducing errors, and improving performance through collaborative inference.

Contribution

It proposes a novel dynamic verifier that actively co-solves tasks with the policy model, enabling real-time error detection and correction during reasoning processes.

Findings

Achieves improved reasoning accuracy on MathVista, MathVerse, and MMMU datasets.

Enables an 8B-parameter model to perform competitively with larger models.

Develops CoRe dataset for training the guided verifier with process-level negatives.

Abstract

Reinforcement Learning (RL) has emerged as a pivotal mechanism for enhancing the complex reasoning capabilities of Multimodal Large Language Models (MLLMs). However, prevailing paradigms typically rely on solitary rollout strategies where the model works alone. This lack of intermediate oversight renders the reasoning process susceptible to error propagation, where early logical deviations cascade into irreversible failures, resulting in noisy optimization signals. In this paper, we propose the \textbf{Guided Verifier} framework to address these structural limitations. Moving beyond passive terminal rewards, we introduce a dynamic verifier that actively co-solves tasks alongside the policy. During the rollout phase, this verifier interacts with the policy model in real-time, detecting inconsistencies and providing directional signals to steer the model toward valid trajectories. To facilitate this, we develop a specialized data synthesis pipeline targeting multimodal hallucinations, constructing \textbf{CoRe} dataset of process-level negatives and \textbf{Co}rrect-guide \textbf{Re}asoning trajectories to train the guided verifier. Extensive experiments on MathVista, MathVerse and MMMU indicate that by allocating compute to collaborative inference and dynamic verification, an 8B-parameter model can achieve strong performance.