DAMA: Data- and Model-aware Alignment of Multi-modal LLMs

Jinda Lu, Junkang Wu, Jinghan Li, Xiaojun Jia, Shuo Wang, YiFan Zhang, Junfeng Fang, Xiang Wang, Xiangnan He

Abstract:

Direct Preference Optimization (DPO) has shown effectiveness in aligning multi-modal large language models (MLLM) with human preferences.
However, existing methods exhibit an imbalanced responsiveness to the data of varying hardness, tending to overfit on the easy-to-distinguish data while underfitting on the hard-to-distinguish data.
In this paper, we propose Data- and Model-aware DPO (DAMA) to dynamically adjust the optimization process from two key aspects: (1) a data-aware strategy that incorporates data hardness, and (2) a model-aware strategy that integrates real-time model responses. By combining the two strategies, DAMA enables the model to effectively adapt to data with varying levels of hardness.
Extensive experiments on five benchmarks demonstrate that DAMA not only significantly enhances the trustworthiness, but also improves the effectiveness over general tasks. For instance, on the Object HalBench, our DAMA-7B reduces response-level and mentioned-level hallucination by 90.0% and 95.3%, respectively, surpassing the performance of GPT-4V.

Lay Summary: Current DPO algorithms for multi-modal LLMs exhibit imbalanced responsiveness to data with various hardness. We propose a data- and model-aware strategy to tackle this, this will help the MLLMs effectively utilize the preference data.

Link To Code: https://github.com/injadlu/DAMA

Primary Area: Applications->Computer Vision

Keywords: Multi-modal Large Language Models, Preference Optimization

Submission Number: 5482