理解并缓解多模态推理链模型中的幻觉
Understanding and Mitigating Hallucinations in Multimodal Chain-of-Thought Models
AI总结：
本文研究多模态推理链模型中的幻觉问题，发现其源于联想推理步骤中的虚假文本生成，提出一种有效策略缓解幻觉，实验表明方法效果显著且可与其他缓解方法结合提升性能。
摘要：
多模态推理链（MCoT）模型在复杂视觉推理任务中表现出色，但近期研究表明其在生成过程中因视觉注意力减弱而产生严重幻觉问题。然而，视觉注意力衰减是大型视觉-语言模型（LVLMs）中已知的问题。考虑到MCoT模型与传统LVLMs在推理过程中的根本差异，我们提出基本问题：MCoT模型是否具有独特的幻觉原因？为回答此问题，我们系统研究了MCoT模型的幻觉模式，发现虚假文本主要在联想推理步骤中生成，我们称之为发散思维。基于这些见解，我们引入一种简单但有效的策略，能够有效定位发散思维步骤并在解码过程中干预以缓解幻觉。大量实验表明，我们的方法在很大程度上优于现有方法。更重要的是，我们提出的方法可以方便地与其他幻觉缓解方法结合，并进一步提升其性能。代码已公开在https://github.com/ASGO-MM/MCoT-hallucination。
英文摘要：
Multimodal Chain-of-Thought (MCoT) models have demonstrated impressive capability in complex visual reasoning tasks. Unfortunately, recent studies reveal that they suffer from severe hallucination problems due to diminished visual attention during the generation process. However, visual attention decay is a well-studied problem in Large Vision-Language Models (LVLMs). Considering the fundamental differences in reasoning processes between MCoT models and traditional LVLMs, we raise a basic question: Whether MCoT models have unique causes of hallucinations? To answer this question, we systematically investigate the hallucination patterns of MCoT models and find that fabricated texts are primarily generated in associative reasoning steps, which we term divergent thinking. Leveraging these insights, we introduce a simple yet effective strategy that can effectively localize divergent thinking steps and intervene in the decoding process to mitigate hallucinations. Extensive experiments show that our method outperforms existing methods by a large margin. More importantly, our proposed method can be conveniently integrated with other hallucination mitigation methods and further boost their performance. The code is publicly available at https://github.com/ASGO-MM/MCoT-hallucination.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
