arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.02028cs.AI
收藏

编辑知识,而不仅仅是事实:基于背景故事的多步推理

Edit Knowledge, Not Just Facts via Multi-Step Reasoning over Background Stories

Aalto University(阿莱大学);Amazon.com(亚马逊公司);System 2 AI(系统2人工智能)

Ya Gao, Kalle Kujanpää, Pekka Marttinen, Harri Valpola, Alexander Ilin

AI总结:

提出将知识更新视为推理问题,通过背景故事引入新知识、自生成多跳问题训练和知识蒸馏,使模型在多步推理中灵活运用新信息。

摘要:

使人工智能系统,特别是大型语言模型,能够更新知识并在推理过程中灵活应用,仍然是一个核心挑战。现有的知识编辑方法强调原子事实,改进了事实回忆,但往往未能将更新后的信息整合到可在不同上下文中使用的连贯框架中。在这项工作中,我们认为知识更新从根本上是一个推理问题,而不是记忆问题。因此,模型应该在新的信息有助于解决任务的情况下进行训练,结合已有知识,并通过多步推理进行练习。基于这一见解,我们提出了一种基于三个原则的训练策略。首先,新知识作为连贯的背景故事引入,将新事实情境化并解释它们与现有知识的关系。其次,使用自生成的多跳问题训练模型,这些问题需要涉及新信息的多步推理。第三,使用知识蒸馏进行训练,迫使学生模型内化教师的推理行为,而无需访问新信息。实验表明,使用此策略训练的模型在推理过程中有效利用新获得的知识,并在需要结合多个新事实的具有挑战性的问题上取得了显著性能。

英文摘要:

Enabling artificial intelligence systems, particularly large language models, to update knowledge and flexibly apply it during reasoning remains a central challenge. Existing knowledge editing approaches emphasize atomic facts, improving factual recall but often failing to integrate updated information into a coherent framework usable across contexts. In this work, we argue that knowledge update is fundamentally a reasoning problem rather than a memorization problem. Consequently, a model should be trained in situations where the new information is instrumental to solving a task, combined with pre-existing knowledge, and exercised through multi-step reasoning. Based on this insight, we propose a training strategy based on three principles. First, new knowledge is introduced as a coherent background story that contextualizes novel facts and explains their relation to existing knowledge. Second, models are trained using self-generated multi-hop questions that require multi-step reasoning involving the new information. Third, training is done using knowledge distillation, forcing a student model to internalize the teacher's reasoning behavior without access to the novel information. Experiments show that models trained with this strategy effectively leverage newly acquired knowledge during reasoning and achieve remarkable performance on challenging questions that require combining multiple new facts.

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。

↑