通过动态调整图像令牌注意力，减少大模型幻觉。
MDSAM:Memory-Driven Sparse Attention Matrix for LVLMs Hallucination Mitigation
- 基于记忆机制动态优化每层图像令牌的注意力分配。
- 在多任务基准上显著降低幻觉率，提升生成可靠性。
- 无需训练即可适配多种模型，适合实际部署场景。
大型视觉语言模型（LVLMs）的幻觉问题常源于解码时对图像令牌的敏感性，表现为生成真实与虚构实体时均出现注意力峰值。为此，我们提出无需训练的内存驱动稀疏注意力矩阵（MDSAM），该方法在每层动态捕捉并优化图像令牌的注意力分配。MDSAM通过解码过程中的对齐激活记忆更新，增强对相关图像令牌的关注，有效抑制幻觉。我们在图像描述生成和视觉问答等多个基准上评估了MDSAM，结果表明其能持续降低幻觉率并提升可靠性。该方法兼容多种LVLM架构，无需额外训练或外部工具，展现出良好的适应性与有效性。
原文摘要 · Abstract (English)
Hallucinations in large vision-language models (LVLMs) often stem from the model's sensitivity to image tokens during decoding, as evidenced by attention peaks observed when generating both real and hallucinated entities. To address this, we propose Memory-Driven Sparse Attention Matrix (MDSAM) , a novel training-free approach that dynamically captures and refines the attention allocated to image tokens at each layer. MDSAM memorizes attention patterns and activates updates through alignment during decoding, enhancing focus on relevant image tokens while effectively reducing hallucinations. We evaluate MDSAM on multiple benchmarks for tasks such as image captioning and visual question answering, demonstrating its ability to consistently reduce hallucinations and improve reliability. Compatible with various LVLM architectures, MDSAM highlights its adaptability and effectiveness in mitigating hallucinations without requiring additional training or external tools.
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。