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 2409.05677cs.CLcs.AIcs.CEcs.ETcs.IR
收藏

RIRAG:监管信息检索与答案生成

RIRAG: Regulatory Information Retrieval and Answer Generation

Mohamed Bin Zayed University of Artificial Intelligence (MBZUAI)(穆罕默德·本·扎耶德人工智能大学(MBZUAI));Technical University of Darmstadt(达姆施塔特工业大学);Hessian Center for AI (hessian.AI)(黑森人工智能中心(hessian.AI))

Tuba Gokhan, Kexin Wang, Iryna Gurevych, Ted Briscoe

AI总结:

本文提出问题—监管段落对生成任务,构建含27,869个问题的 ObliQA 数据集,设计基线 RIRAG 系统,并用新型指标 RePASs 评估答案的义务覆盖与矛盾情况。

摘要:

由政府监管机构发布的监管文件确立了各组织为实现法律合规而必须遵守的规则、指南和标准。这些文件以篇幅长、复杂性高且更新频繁为特征,解释起来颇具挑战,需要各组织投入大量时间和专业知识以确保持续合规。监管自然语言处理(RegNLP)是一个多学科领域,旨在简化监管规则与义务的获取和解释。我们引入了一项生成问题—段落对的任务,即自动创建问题并将其与相关监管段落配对,以促进监管问答系统的开发。我们构建了 ObliQA 数据集,其中包含源自阿布扎比全球市场(ADGM)金融监管文档集合的 27,869 个问题;设计了一个基线监管信息检索与答案生成(RIRAG)系统,并使用 RePASs 对其进行评估。RePASs 是一种新型评估指标,用于检验生成的答案是否准确涵盖所有相关义务,同时避免矛盾。

英文摘要:

Regulatory documents, issued by governmental regulatory bodies, establish rules, guidelines, and standards that organizations must adhere to for legal compliance. These documents, characterized by their length, complexity and frequent updates, are challenging to interpret, requiring significant allocation of time and expertise on the part of organizations to ensure ongoing compliance. Regulatory Natural Language Processing (RegNLP) is a multidisciplinary field aimed at simplifying access to and interpretation of regulatory rules and obligations. We introduce a task of generating question-passages pairs, where questions are automatically created and paired with relevant regulatory passages, facilitating the development of regulatory question-answering systems. We create the ObliQA dataset, containing 27,869 questions derived from the collection of Abu Dhabi Global Markets (ADGM) financial regulation documents, design a baseline Regulatory Information Retrieval and Answer Generation (RIRAG) system and evaluate it with RePASs, a novel evaluation metric that tests whether generated answers accurately capture all relevant obligations while avoiding contradictions.

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。

↑