Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Weiyun, Chen, Zhe, Wang, Wenhai, Cao, Yue, Liu, Yangzhou, Gao, Zhangwei, Zhu, Jinguo, Zhu, Xizhou, Lu, Lewei, Qiao, Yu, Dai, Jifeng
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912312501207040
author Wang, Weiyun
Chen, Zhe
Wang, Wenhai
Cao, Yue
Liu, Yangzhou
Gao, Zhangwei
Zhu, Jinguo
Zhu, Xizhou
Lu, Lewei
Qiao, Yu
Dai, Jifeng
author_facet Wang, Weiyun
Chen, Zhe
Wang, Wenhai
Cao, Yue
Liu, Yangzhou
Gao, Zhangwei
Zhu, Jinguo
Zhu, Xizhou
Lu, Lewei
Qiao, Yu
Dai, Jifeng
contents Existing open-source multimodal large language models (MLLMs) generally follow a training process involving pre-training and supervised fine-tuning. However, these models suffer from distribution shifts, which limit their multimodal reasoning, particularly in the Chain-of-Thought (CoT) performance. To address this, we introduce a preference optimization (PO) process to enhance the multimodal reasoning capabilities of MLLMs. Specifically, (1) on the data side, we design an automated preference data construction pipeline to create MMPR, a high-quality, large-scale multimodal reasoning preference dataset; and (2) on the model side, we explore integrating PO with MLLMs, developing a simple yet effective method, termed Mixed Preference Optimization (MPO), which boosts multimodal CoT performance. Our approach enhances the multimodal reasoning abilities of both InternVL2-8B and InternVL2-76B. Notably, our model, InternVL2-8B-MPO, achieves an accuracy of 67.0 on MathVista, outperforming InternVL2-8B by 8.7 points and achieving performance comparable to the 10$\times$ larger InternVL2-76B. We hope this study could inspire further advancements in MLLMs. Code, data, and model are released.
format Preprint
id arxiv_https___arxiv_org_abs_2411_10442
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization
Wang, Weiyun
Chen, Zhe
Wang, Wenhai
Cao, Yue
Liu, Yangzhou
Gao, Zhangwei
Zhu, Jinguo
Zhu, Xizhou
Lu, Lewei
Qiao, Yu
Dai, Jifeng
Computation and Language
Computer Vision and Pattern Recognition
Existing open-source multimodal large language models (MLLMs) generally follow a training process involving pre-training and supervised fine-tuning. However, these models suffer from distribution shifts, which limit their multimodal reasoning, particularly in the Chain-of-Thought (CoT) performance. To address this, we introduce a preference optimization (PO) process to enhance the multimodal reasoning capabilities of MLLMs. Specifically, (1) on the data side, we design an automated preference data construction pipeline to create MMPR, a high-quality, large-scale multimodal reasoning preference dataset; and (2) on the model side, we explore integrating PO with MLLMs, developing a simple yet effective method, termed Mixed Preference Optimization (MPO), which boosts multimodal CoT performance. Our approach enhances the multimodal reasoning abilities of both InternVL2-8B and InternVL2-76B. Notably, our model, InternVL2-8B-MPO, achieves an accuracy of 67.0 on MathVista, outperforming InternVL2-8B by 8.7 points and achieving performance comparable to the 10$\times$ larger InternVL2-76B. We hope this study could inspire further advancements in MLLMs. Code, data, and model are released.
title Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization
topic Computation and Language
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.10442