arXiv 2410.152672026-07-17cs.CRcs.CL

当模型遗忘遇上检索增强生成（RAG）：保守秘密还是遗忘知识？

When Machine Unlearning Meets Retrieval-Augmented Generation (RAG): Keep Secret or Forget Knowledge?

School of Computer Science, University of Technology Sydney（悉尼科技大学计算机科学学院）；Faculty of Data Science, City University of Macau（澳门城市大学数据科学学院）

Shang Wang, Tianqing Zhu, Dayong Ye, Wanlei Zhou

备注：: This paper is accepted by IEEE Transactions on Dependable and Secure Computing 2025. The source code is available at \url{https://github.com/shihe98/RAG_Unlearning}

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

研究大语言模型训练中敏感信息保留问题，提出基于检索增强生成（RAG）技术的轻量级行为遗忘框架，通过修改外部知识库模拟遗忘，经实验验证该方法符合五个关键遗忘标准，还可扩展到多模态模型和智能体。

摘要：

ChatGPT和Gemini等大语言模型（LLMs）展现出强大的自然语言生成能力，但训练中会无意间学习并保留敏感信息和有害内容，引发伦理和法律问题。为解决这些问题，引入了模型遗忘。现有遗忘方法存在计算需求高、适用性有限或灾难性遗忘风险等问题。我们提出基于检索增强生成（RAG）技术的轻量级行为遗忘框架，通过修改RAG的外部知识库模拟遗忘效果，将构建遗忘知识视为约束优化问题，得出两个关键组件。该方法对闭源LLMs特别有效，现有方法常在此失败。通过在开源和闭源模型上的大量实验评估框架，结果表明该方法符合有效性、通用性、无害性、简单性和鲁棒性五个关键遗忘标准，且可扩展到多模态大语言模型和基于LLM的智能体。

英文摘要：

The deployment of large language models (LLMs) like ChatGPT and Gemini has shown their powerful natural language generation capabilities. However, these models can inadvertently learn and retain sensitive information and harmful content during training, raising significant ethical and legal concerns. To address these issues, machine unlearning has been introduced as a potential solution. While existing unlearning methods take into account the specific characteristics of LLMs, they often suffer from high computational demands, limited applicability, or the risk of catastrophic forgetting. To address these limitations, we propose a lightweight behavioral unlearning framework based on Retrieval-Augmented Generation (RAG) technology. By modifying the external knowledge base of RAG, we simulate the effects of forgetting without directly interacting with the unlearned LLM. We approach the construction of unlearned knowledge as a constrained optimization problem, deriving two key components that underpin the effectiveness of RAG-based unlearning. This RAG-based approach is particularly effective for closed-source LLMs, where existing unlearning methods often fail. We evaluate our framework through extensive experiments on both open-source and closed-source models, including ChatGPT, Gemini, Llama-2-7b-chat, and PaLM 2. The results demonstrate that our approach meets five key unlearning criteria: effectiveness, universality, harmlessness, simplicity, and robustness. Meanwhile, this approach can extend to multimodal large language models and LLM-based agents.

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。