让多模态大模型学会自我反思式知识更新。
Towards Meta-Cognitive Knowledge Editing for Multimodal LLMs
- 构建元认知记忆与博弈机制，让模型自知知识是否正确。
- 在三种新评测中表现优于现有方法，尤其抗噪声能力强。
- 适合研究大模型自我修正与可信知识管理的学者。
知识编辑使多模态大语言模型能高效更新过时或错误信息。然而，现有基准主要关注认知层面的修改，缺乏对深层元认知过程的关注。为此，我们提出CogEdit，一个新型基准，用于评估多模态大模型在三个层次上的元认知知识编辑能力：(1) 反事实驱动编辑，评估模型对知识正确性变化的自我意识；(2) 边界约束编辑，确保泛化合理且无意外干扰；(3) 噪声鲁棒编辑，促进对不确定信息的反思性评估。为推进元认知编辑，我们提出MIND（Meta-cognitive INtegrated Dynamic Knowledge Editing）框架，该框架构建元知识记忆以实现自我意识，采用博弈论交互监控知识激活，并引入标签精炼实现抗噪声更新。大量实验表明，MIND显著优于现有认知编辑方法，在传统与元认知知识编辑基准上均表现优异。
原文摘要 · Abstract (English)
Knowledge editing enables multimodal large language models (MLLMs) to efficiently update outdated or incorrect information. However, existing benchmarks primarily emphasize cognitive-level modifications while lacking a focus on deeper meta-cognitive processes. To bridge this gap, we introduce CogEdit, a novel benchmark designed to evaluate MLLMs' meta-cognitive knowledge editing abilities across three levels: (1) Counterfactual-Driven Editing, assessing self-awareness of knowledge correctness changes; (2) Boundary Constraint Editing, ensuring appropriate generalization without unintended interference; and (3) Noise-Robust Editing, promoting reflective evaluation of uncertain information. To advance meta-cognitive editing, we propose MIND (Meta-cognitive INtegrated Dynamic Knowledge Editing), a framework that constructs a meta-knowledge memory for self-awareness, employs game-theoretic interactions to monitor knowledge activation, and incorporates label refinement for noise-robust updates. Extensive experiments show that MIND significantly outperforms existing cognitive editing approaches, achieving strong performance on both traditional and meta-cognitive knowledge editing benchmarks.
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。