Improve Vision Language Model Chain-of-thought Reasoning

TL;DR

This paper enhances vision language models' reasoning by enriching training data with GPT-4o rationales and applying reinforcement learning, leading to improved interpretability and generalization in reasoning tasks.

Contribution

It introduces a method to incorporate detailed rationales and reinforcement learning to significantly improve CoT reasoning in vision language models.

Findings

Improved CoT reasoning performance on benchmark datasets

Enhanced generalization to direct answer prediction

Effective use of GPT-4o generated rationales and reinforcement learning

Abstract

Chain-of-thought (CoT) reasoning in vision language models (VLMs) is crucial for improving interpretability and trustworthiness. However, current training recipes lack robust CoT reasoning data, relying on datasets dominated by short annotations with minimal rationales. In this work, we show that training VLM on short answers does not generalize well to reasoning tasks that require more detailed responses. To address this, we propose a two-fold approach. First, we distill rationales from GPT-4o model to enrich the training data and fine-tune VLMs, boosting their CoT performance. Second, we apply reinforcement learning to further calibrate reasoning quality. Specifically, we construct positive (correct) and negative (incorrect) pairs of model-generated reasoning chains, by comparing their predictions with annotated short answers. Using this pairwise data, we apply the Direct Preference Optimization algorithm to refine the model's reasoning abilities. Our experiments demonstrate significant improvements in CoT reasoning on benchmark datasets and better generalization to direct answer prediction as well. This work emphasizes the importance of incorporating detailed rationales in training and leveraging reinforcement learning to strengthen the reasoning capabilities of VLMs.