Multimodal large language models (MLLMs) contribute a powerful mechanism to understanding visual information building on large language models. However, MLLMs are notorious for suffering from hallucinations, especially when generating lengthy, detailed descriptions for images.
Our analysis reveals that hallucinations stem from the inherent summarization mechanism of large language models, leading to excessive dependence on linguistic tokens while neglecting vision information. In this paper, we propose NoiseBoost, a broadly applicable and simple method for alleviating hallucinations for MLLMs through the integration of noise feature perturbations. Noise perturbation acts as a regularizer, facilitating a balanced distribution of attention weights among visual and linguistic tokens.
Despite its simplicity, NoiseBoost consistently enhances the performance of MLLMs across common training strategies, including supervised fine-tuning and reinforcement learning. Further, NoiseBoost pioneerly enables semi-supervised learning for MLLMs, unleashing the power of unlabeled data. Comprehensive experiments demonstrate that NoiseBoost improves dense caption accuracy by 8.1\% with human evaluation and achieves comparable results with 50\% of the data by mining unlabeled data. The code and data will be made publicly accessible.
@article{wu2024noiseboost,
title={NoiseBoost: Alleviating Hallucination with Noise Perturbation for Multimodal Large Language Models},
author={Wu, Kai and Jiang, Boyuan and Jiang, Zhengkai and He, Qingdong and Luo, Donghao and Wang, Shengzhi and Liu, Qingwen and Wang, Chengjie},
journal={arXiv preprint arXiv:2405.20081},
year={2024}
}