通过指令对比定位干扰视觉元素，减少大模型幻觉
Instruction-Aligned Visual Attention for Mitigating Hallucinations in Large Vision-Language Models
- 用不同指令下注意力变化识别无关图像片段
- 对比解码动态调整无关区域的输出概率
- 在多个数据集上显著降低物体幻觉率
尽管大型视觉语言模型（LVLMs）取得了显著进展，但在描述图像时仍会出现幻觉，生成并不存在的物体。研究发现，这些模型倾向于过度关注某些与问题无关的图像标记，从而扭曲输出。为此，本文提出指令对齐视觉注意力（IAVA）方法，通过比较两种不同指令下的注意力权重变化，识别出无关图像标记。利用对比解码，动态调整原始图像标记与无关图像标记产生的逻辑值，减少模型对无关信息的过度关注。实验结果表明，IAVA在MME、POPE和TextVQA等基准测试中，持续优于现有解码技术，在缓解物体幻觉方面表现更优。代码已公开于https://github.com/Lee-lab558/IAVA。
原文摘要 · Abstract (English)
Despite the significant success of Large Vision-Language models(LVLMs), these models still suffer hallucinations when describing images, generating answers that include non-existent objects. It is reported that these models tend to over-focus on certain irrelevant image tokens that do not contain critical information for answering the question and distort the output. To address this, we propose an Instruction-Aligned Visual Attention(IAVA) approach, which identifies irrelevant tokens by comparing changes in attention weights under two different instructions. By applying contrastive decoding, we dynamically adjust the logits generated from original image tokens and irrelevant image tokens, reducing the model's over-attention to irrelevant information. The experimental results demonstrate that IAVA consistently outperforms existing decoding techniques on benchmarks such as MME, POPE, and TextVQA in mitigating object hallucinations. Our IAVA approach is available online at https://github.com/Lee-lab558/IAVA.
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。