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 2311.09358cs.CLcs.AI
收藏

检索增强语言模型在科学任务中不确定性量化的实证评估

Empirical evaluation of Uncertainty Quantification in Retrieval-Augmented Language Models for Science

Pacific Northwest National Laboratory(太平洋西北国家实验室)

Sridevi Wagle, Sai Munikoti, Anurag Acharya, Sara Smith, Sameera Horawalavithana

AI总结:

本研究实证评估了检索增强语言模型在科学任务中的不确定性量化,发现模型过度自信,且科学知识无法缓解此问题。

摘要:

大型语言模型(LLMs)在自然语言处理任务中展现了卓越的成就,能够生成高质量的输出。然而,LLMs仍然存在局限性,包括生成事实错误的信息。在安全关键的应用中,评估LLM生成内容的置信度以做出明智决策非常重要。检索增强语言模型(RALMs)是自然语言处理中一个相对较新的研究领域。RALMs为科学NLP任务提供了潜在的好处,因为检索到的文档可以作为证据来支持模型生成的内容。这种证据的加入增强了可信度,因为用户可以验证和探索检索到的文档以确认模型输出。量化RALM生成中的不确定性进一步提高了可信度,检索到的文本和置信度分数有助于为科学应用构建一个全面且可靠的模型。然而,关于RALMs的不确定性量化(UQ)研究很少甚至没有,特别是在科学背景下。本研究旨在通过针对科学任务对RALMs中的UQ进行全面评估来填补这一空白。本研究调查了当科学知识作为预训练和检索数据纳入时,不确定性分数如何变化,并探讨了不确定性分数与模型生成输出准确性之间的关系。我们观察到,一个现有的以科学知识作为检索数据微调的RALM在生成预测时往往比仅用科学知识预训练的模型更自信。我们还发现,RALMs对其预测过度自信,使得不准确的预测比准确的预测更自信。无论是作为预训练还是检索语料库提供的科学知识都不能帮助缓解这个问题。我们在https://github.com/pnnl/EXPERT2上发布了我们的代码、数据和仪表板。

英文摘要:

Large language models (LLMs) have shown remarkable achievements in natural language processing tasks, producing high-quality outputs. However, LLMs still exhibit limitations, including the generation of factually incorrect information. In safety-critical applications, it is important to assess the confidence of LLM-generated content to make informed decisions. Retrieval Augmented Language Models (RALMs) is relatively a new area of research in NLP. RALMs offer potential benefits for scientific NLP tasks, as retrieved documents, can serve as evidence to support model-generated content. This inclusion of evidence enhances trustworthiness, as users can verify and explore the retrieved documents to validate model outputs. Quantifying uncertainty in RALM generations further improves trustworthiness, with retrieved text and confidence scores contributing to a comprehensive and reliable model for scientific applications. However, there is limited to no research on UQ for RALMs, particularly in scientific contexts. This study aims to address this gap by conducting a comprehensive evaluation of UQ in RALMs, focusing on scientific tasks. This research investigates how uncertainty scores vary when scientific knowledge is incorporated as pretraining and retrieval data and explores the relationship between uncertainty scores and the accuracy of model-generated outputs. We observe that an existing RALM finetuned with scientific knowledge as the retrieval data tends to be more confident in generating predictions compared to the model pretrained only with scientific knowledge. We also found that RALMs are overconfident in their predictions, making inaccurate predictions more confidently than accurate ones. Scientific knowledge provided either as pretraining or retrieval corpus does not help alleviate this issue. We released our code, data and dashboards at https://github.com/pnnl/EXPERT2.

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。

↑