上下文编辑：从自我诱导分布中学习知识

In-Context Editing: Learning Knowledge from Self-Induced Distributions

June 17, 2024

作者: Siyuan Qi, Bangcheng Yang, Kailin Jiang, Xiaobo Wang, Jiaqi Li, Yifan Zhong, Yaodong Yang, Zilong Zheng

cs.AI

摘要

现有的语言模型微调范式在知识编辑场景中表现脆弱，即模型必须在无需进行大量重新训练的情况下整合新信息。这种脆弱性通常导致过拟合、性能降低和生成不自然语言。为了解决这一问题，我们提出了一种名为一致上下文编辑（ICE）的新方法，利用模型的上下文学习能力朝向上下文分布进行调整，而非单热目标。ICE引入了一个简单的优化框架，包括目标和过程，增强了基于梯度调整方法的鲁棒性和有效性。我们从知识编辑的四个关键方面：准确性、局部性、泛化性和语言质量，提供了ICE的分析见解，展示了其优势。在四个数据集上的实验结果证实了ICE的有效性，并展示了其在持续编辑方面的潜力，确保更新的信息被整合同时保持模型的完整性。

English

The existing fine-tuning paradigm for language models is brittle in knowledge editing scenarios, where the model must incorporate new information without extensive retraining. This brittleness often results in overfitting, reduced performance, and unnatural language generation. To address this, we propose Consistent In-Context Editing (ICE), a novel approach that leverages the model's in-context learning capability to tune toward a contextual distribution rather than a one-hot target. ICE introduces a straightforward optimization framework that includes both a target and a procedure, enhancing the robustness and effectiveness of gradient-based tuning methods. We provide analytical insights into ICE across four critical aspects of knowledge editing: accuracy, locality, generalization, and linguistic quality, showing its advantages. Experimental results across four datasets confirm the effectiveness of ICE and demonstrate its potential for continual editing, ensuring that updated information is incorporated while preserving the integrity of the model.