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 2404.08417cs.LGcs.AIcs.CL
收藏

AdapterSwap:具备数据移除与访问控制保障的大语言模型持续训练

AdapterSwap: Continuous Training of LLMs with Data Removal and Access-Control Guarantees

Johns Hopkins University(约翰斯·霍普金斯大学)

William Fleshman, Aleem Khan, Marc Marone, Benjamin Van Durme

AI总结:

AdapterSwap通过将知识组织为低秩适配器并在推理时动态组合,实现了大语言模型的高效持续学习,同时支持细粒度的数据访问控制与删除保证。

摘要:

大语言模型(LLMs)越来越能够通过从静态预训练语料库中回忆信息来完成知识密集型任务。本文关注的是在数据需求不断演变背景下的LLMs。例如:周期性引入的新数据批次;具有基于用户的访问控制的数据子集;或对文档动态移除的需求,并保证相关知识无法被回忆。我们希望满足这些需求,同时确保模型在新数据可用时不会遗忘旧信息。为了解决这些问题,我们提出了AdapterSwap,这是一种训练与推理方案,将来自数据集合的知识组织为一组低秩适配器,并在推理过程中动态组合这些适配器。我们的实验证明了AdapterSwap支持高效持续学习的能力,同时也使组织能够对数据访问和删除进行细粒度控制。

英文摘要:

Large language models (LLMs) are increasingly capable of completing knowledge intensive tasks by recalling information from a static pretraining corpus. Here we are concerned with LLMs in the context of evolving data requirements. For instance: batches of new data that are introduced periodically; subsets of data with user-based access controls; or requirements on dynamic removal of documents with guarantees that associated knowledge cannot be recalled. We wish to satisfy these requirements while at the same time ensuring a model does not forget old information when new data becomes available. To address these issues, we introduce AdapterSwap, a training and inference scheme that organizes knowledge from a data collection into a set of low-rank adapters, which are dynamically composed during inference. Our experiments demonstrate AdapterSwap's ability to support efficient continual learning, while also enabling organizations to have fine-grained control over data access and deletion.

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。

↑