发现多模态推理模型幻觉源于联想步骤，提出干预解码有效抑制。
Understanding and Mitigating Hallucinations in Multimodal Chain-of-Thought Models
- 识别出幻觉主要发生在联想推理步骤，称为发散思维。
- 所提方法显著降低幻觉率，性能优于现有技术。
- 可无缝集成其他消幻方法，适合视觉推理研究者使用。
多模态链式思维（MCoT）模型在复杂视觉推理任务中表现优异，但近期研究发现其因生成过程中的视觉注意力减弱而存在严重幻觉问题。尽管视觉注意力衰减在大型视觉语言模型（LVLMs）中已有深入研究，但MCoT模型与传统LVLM在推理机制上存在根本差异，因此我们提出核心问题：MCoT模型的幻觉是否具有独特成因？通过系统分析MCoT模型的幻觉模式，我们发现虚构文本主要产生于联想推理步骤，称之为发散思维。基于此洞察，我们提出一种简单而有效的策略，可精准定位发散思维步骤并干预解码过程以缓解幻觉。大量实验表明，该方法显著优于现有方法；更重要的是，该方法可便捷地与其它消幻技术结合，进一步提升其性能。代码已公开于 https://github.com/ASGO-MM/MCoT-hallucination。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。