mDPO: Conditional Preference Optimization for Multimodal Large Language Models
This work addresses alignment challenges for multimodal large language models, offering a domain-specific improvement in reducing hallucination.
The paper tackled the unconditional preference problem in multimodal preference optimization, where models overlook image conditions, by proposing mDPO, which optimizes image preference and uses a reward anchor to improve performance, significantly reducing hallucination in experiments on two models and three benchmarks.
Direct preference optimization (DPO) has shown to be an effective method for large language model (LLM) alignment. Recent works have attempted to apply DPO to multimodal scenarios but have found it challenging to achieve consistent improvement. Through a comparative experiment, we identify the unconditional preference problem in multimodal preference optimization, where the model overlooks the image condition. To address this problem, we propose mDPO, a multimodal DPO objective that prevents the over-prioritization of language-only preferences by also optimizing image preference. Moreover, we introduce a reward anchor that forces the reward to be positive for chosen responses, thereby avoiding the decrease in their likelihood -- an intrinsic problem of relative preference optimization. Experiments on two multimodal LLMs of different sizes and three widely used benchmarks demonstrate that mDPO effectively addresses the unconditional preference problem in multimodal preference optimization and significantly improves model performance, particularly in reducing hallucination.