面向多跳问答的语言模型检索增强知识编辑
Retrieval-enhanced Knowledge Editing in Language Models for Multi-Hop Question Answering
AI总结：
针对大语言模型多跳问答中知识更新难的问题，提出检索增强模型编辑框架RAE，通过互信息最大化检索链式事实、剪枝冗余信息，经多模型验证可提升知识更新后的问答准确率。
摘要：
大语言模型（LLMs）在问答任务中表现出色，但往往难以整合实时知识，导致回复可能过时或不准确。这一问题在处理多跳问题时更具挑战性，因为这类问题要求LLMs更新并整合与问题相关的多条知识。为解决该问题，我们提出面向多跳问答的检索增强模型编辑（Retrieval-Augmented model Editing, RAE）框架。RAE首先检索已编辑的事实，再通过上下文学习优化语言模型。具体而言，我们基于互信息最大化的检索方法利用LLMs的推理能力，识别传统基于相似度的搜索可能遗漏的链式事实。此外，该框架包含一种剪枝策略，用于去除检索事实中的冗余信息，提升编辑准确率并缓解幻觉问题。我们的框架在事实检索有效性上有理论依据支撑。最后，在多种LLMs上开展的综合评估验证了RAE能够基于更新后的知识给出准确答案。我们的代码可在https://github.com/sycny/RAE获取。
英文摘要：
Large Language Models (LLMs) have shown proficiency in question-answering tasks but often struggle to integrate real-time knowledge, leading to potentially outdated or inaccurate responses. This problem becomes even more challenging when dealing with multi-hop questions, since they require LLMs to update and integrate multiple knowledge pieces relevant to the questions. To tackle the problem, we propose the Retrieval-Augmented model Editing (RAE) framework for multi-hop question answering. RAE first retrieves edited facts and then refines the language model through in-context learning. Specifically, our retrieval approach, based on mutual information maximization, leverages the reasoning abilities of LLMs to identify chain facts that traditional similarity-based searches might miss. In addition, our framework includes a pruning strategy to eliminate redundant information from the retrieved facts, which enhances the editing accuracy and mitigates the hallucination problem. Our framework is supported by theoretical justification for its fact retrieval efficacy. Finally, comprehensive evaluation across various LLMs validates RAE's ability in providing accurate answers with updated knowledge. Our code is available at: https://github.com/sycny/RAE.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
