通过图像引导的注意力头抑制，显著降低视觉语言模型幻觉。
Mitigating Hallucinations in Vision-Language Models through Image-Guided Head Suppression
- 识别并抑制低关注图像的注意力头，保留高相关性头。
- 幻觉率降低2.7倍，推理速度提升1.8倍。
- 无需额外计算，适合实时应用的轻量级方法。
尽管大型视觉语言模型（LVLM）在多模态理解任务中取得显著进展，但常出现与视觉上下文不符的“幻觉”现象。现有通过推理时干预减少幻觉的方法会显著增加延迟。为此，我们提出SPIN，一种任务无关的注意力引导头抑制策略，可无缝集成于推理过程，几乎不增加计算或延迟开销。我们探究了幻觉是否与特定模型组件相关，分析表明幻觉主要源于每层中动态的注意力头子集。针对每个文本查询词元，我们选择性抑制对图像词元关注度低的注意力头，保留前K个最相关的头。在视觉问答和图像描述任务上的广泛评估表明，SPIN将幻觉分数降低最多2.7倍，同时保持F1性能，并使吞吐量相比现有方法提升1.8倍。代码已公开于https://github.com/YUECHE77/SPIN。
原文摘要 · Abstract (English)
Despite their remarkable progress in multimodal understanding tasks, large vision language models (LVLMs) often suffer from "hallucinations", generating texts misaligned with the visual context. Existing methods aimed at reducing hallucinations through inference time intervention incur a significant increase in latency. To mitigate this, we present SPIN, a task-agnostic attention-guided head suppression strategy that can be seamlessly integrated during inference, without incurring any significant compute or latency overhead. We investigate whether hallucination in LVLMs can be linked to specific model components. Our analysis suggests that hallucinations can be attributed to a dynamic subset of attention heads in each layer. Leveraging this insight, for each text query token, we selectively suppress attention heads that exhibit low attention to image tokens, keeping the top-K attention heads intact. Extensive evaluations on visual question answering and image description tasks demonstrate the efficacy of SPIN in reducing hallucination scores up to 2.7x while maintaining F1, and improving throughput by 1.8x compared to existing alternatives. Code is available at https://github.com/YUECHE77/SPIN.
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。