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 2310.07579cs.LGcs.AIcs.CR
收藏

上下文遗忘：语言模型作为少样本遗忘器

In-Context Unlearning: Language Models as Few Shot Unlearners

Harvard University（哈佛大学）

Martin Pawelczyk, Seth Neel, Himabindu Lakkaraju

备注：
Accepted at ICML 2024
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出上下文遗忘方法，在推理时通过提供带有错误标签的特定输入使大型语言模型无需更新参数即可遗忘训练实例，实验表明其效果可与需要访问模型参数的先进方法相当或更优。

摘要：

机器遗忘研究如何高效地消除特定训练实例对模型的影响，近年来由于“被遗忘权”等监管准则而受到越来越多的关注。实现精确遗忘通常需要对模型进行完整重训练，而对于大型语言模型（LLMs）等超大规模模型而言，这在计算上是不可行的。为此，近期工作提出了若干算法，在不重训练模型的情况下近似移除训练数据。这些算法关键依赖于访问模型参数以进行更新，但在实际中，由于计算限制或只能对LLMs进行查询访问，这一假设可能并不成立。本文提出了一类新的面向LLMs的遗忘方法，称为“上下文遗忘（In-Context Unlearning）”。该方法只需在上下文中提供特定类型的输入，即可使模型遗忘实例，而无需更新模型参数。为了遗忘特定训练实例，我们在推理时将其实例连同与其真实标签不同的标签一起呈现给LLMs。实验结果表明，上下文遗忘与需要访问模型参数的其他最先进方法表现相当，在某些情况下甚至更优，能够在保持测试准确率的同时有效消除特定实例对模型的影响。

英文摘要：

Machine unlearning, the study of efficiently removing the impact of specific training instances on a model, has garnered increased attention in recent years due to regulatory guidelines such as the \emph{Right to be Forgotten}. Achieving precise unlearning typically involves fully retraining the model and is computationally infeasible in case of very large models such as Large Language Models (LLMs). To this end, recent work has proposed several algorithms which approximate the removal of training data without retraining the model. These algorithms crucially rely on access to the model parameters in order to update them, an assumption that may not hold in practice due to computational constraints or having only query access to the LLMs. In this work, we propose a new class of unlearning methods for LLMs called ``In-Context Unlearning.'' This method unlearns instances from the model by simply providing specific kinds of inputs in context, without the need to update model parameters. To unlearn specific training instances, we present these instances to the LLMs at inference time along with labels that differ from their ground truth. Our experimental results demonstrate that in-context unlearning performs on par with, or in some cases outperforms other state-of-the-art methods that require access to model parameters, effectively removing the influence of specific instances on the model while preserving test accuracy.

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。

↑