arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.20111cs.CL
收藏

让大语言模型成为更好的数据创造者

Making Large Language Models Better Data Creators

Information Sciences Institute, University of Southern California(南加州大学信息科学研究所);Microsoft Research(微软研究院)

Dong-Ho Lee, Jay Pujara, Mohit Sewak, Ryen W. White, Sujay Kumar Jauhar

AI总结:

本文提出仅需一个格式示例的统一数据创建流程,证明指令跟随大语言模型可低成本生成高质量训练数据,使下游模型在分布外评估中最高提升17.5%。

摘要:

尽管大语言模型(LLMs)已显著推动自然语言处理(NLP)领域达到最先进水平,但由于成本、响应速度、可控性以及隐私和安全方面的顾虑,将其部署到下游应用中仍具挑战性。因此,可训练模型在某些情况下仍是首选方案。然而,这些模型仍需要人工标注数据才能达到最佳性能,而获取这类数据既昂贵又耗时。为解决这一问题,若干减少人工工作量的技术采用LLMs进行数据标注或数据生成。尽管这些方法在特定应用中有效,但在实际场景中会遇到困难:数据标注需要谨慎的数据选择,而数据生成则需要针对特定任务的提示工程。本文提出一个统一的数据创建流程,该流程仅需一个格式示例,适用于广泛的任务,包括传统上具有语义空标签空间的困难任务。在实验中,我们证明遵循指令的LLMs是极具成本效益的数据创造者,使用这些数据训练的模型在分布外评估中表现出优于使用人工标注数据训练的模型的性能(最高提升17.5%),同时在分布内任务上保持可比性能。这些结果对部署在现实世界中的NLP系统的鲁棒性具有重要意义。

英文摘要:

Although large language models (LLMs) have advanced the state-of-the-art in NLP significantly, deploying them for downstream applications is still challenging due to cost, responsiveness, control, or concerns around privacy and security. As such, trainable models are still the preferred option in some cases. However, these models still require human-labeled data for optimal performance, which is expensive and time-consuming to obtain. In order to address this issue, several techniques to reduce human effort involve labeling or generating data using LLMs. Although these methods are effective for certain applications, in practice they encounter difficulties in real-world scenarios. Labeling data requires careful data selection, while generating data necessitates task-specific prompt engineering. In this paper, we propose a unified data creation pipeline that requires only a single formatting example, and which is applicable to a broad range of tasks, including traditionally problematic ones with semantically devoid label spaces. In our experiments we demonstrate that instruction-following LLMs are highly cost-effective data creators, and that models trained with these data exhibit performance better than those trained with human-labeled data (by up to 17.5%) on out-of-distribution evaluation, while maintaining comparable performance on in-distribution tasks. These results have important implications for the robustness of NLP systems deployed in the real-world.

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。

↑