M3PO: Multimodal-Model-Guided Preference Optimization for Visual Instruction Following
This work addresses the problem of data-efficient preference optimization for LVLMs, which is crucial for developers and researchers in multimodal AI, though it appears incremental as it builds on existing DPO methods with a novel selection mechanism.
The paper tackles the challenge of high-cost human annotation and inefficient sample selection for fine-tuning large vision-language models (LVLMs) in visual instruction following, proposing M3PO, a method that selects high-quality preference pairs from model-generated candidates using a novel scoring mechanism, and demonstrates consistent outperformance over baselines across multiple benchmarks.
Large Vision-Language Models (LVLMs) hold immense potential for complex multimodal instruction following, yet their development is often hindered by the high cost and inconsistency of human annotation required for effective fine-tuning and preference alignment. Traditional supervised fine-tuning (SFT) and existing preference optimization methods like RLHF and DPO frequently struggle to efficiently leverage the model's own generation space to identify highly informative "hard negative" samples. To address these challenges, we propose Multimodal-Model-Guided Preference Optimization (M3PO), a novel and data-efficient method designed to enhance LVLMs' capabilities in visual instruction following. M3PO intelligently selects the most "learning-valuable" preference sample pairs from a diverse pool of LVLM-generated candidates. This selection is driven by a sophisticated mechanism that integrates two crucial signals: a Multimodal Alignment Score (MAS) to assess external quality and the model's Self-Consistency / Confidence (log-probability) to gauge internal belief. These are combined into a novel M3P-Score, which specifically identifies preferred responses and challenging dispreferred responses that the model might confidently generate despite being incorrect. These high-quality preference pairs are then used for efficient Direct Preference Optimization (DPO) fine-tuning on base LVLMs like LLaVA-1.5 (7B/13B) using LoRA. Our extensive experiments demonstrate that M3PO consistently outperforms strong baselines, including SFT, simulated RLHF, vanilla DPO, and RM-DPO, across a comprehensive suite of multimodal instruction following benchmarks (MME-Bench, POPE, IFT, Human Pref. Score).