[论文] LLM-Based Examination of Eligibility Criteria from Securities Prospect...

论文概要

研究领域: NLP 作者: Serhii Hamotskyi, Akash Kumar Gautam, Christian Hanig 发布时间: 2026-06-25 arXiv: 2606.27316

中文摘要

验证证券作为抵押品的资格是德国央行的核心职责之一。然而，在冗长、半结构化且常为双语的招股说明书中，人工核对这些资产是否符合法律和金融标准是一项资源密集型任务。以往工作采用传统的命名实体识别（NER）进行信息提取，但这些方法难以应对 OCR 噪声、语言变异和严格的基于跨度的约束，且每类标注都需要人工标注的训练数据。本文首次呈现将大型语言模型（LLM）应用于资格审查流程的案例研究，将范式转向生成式信息提取流水线。我们的方法将任务分解为提取、归一化和解释三个阶段，在处理噪声文本和德英交错内容时具有更大灵活性。我们进一步引入基于 LLM-as-a-judge 的值评估方法，比基于位置的指标提供更语义化的评估。结果表明，LLM 系统在文档级资格审查上达到高达 91% 的精确率，表现出保守的操作特征，最大程度减少误接受。

原文摘要

Verifying the eligibility of securities as collateral is a key responsibility of the German Central Bank. However, manually verifying these assets against legal and financial criteria within lengthy, semi-structured, and often bilingual prospectuses is a resource-intensive task. While previous efforts utilized traditional Named Entity Recognition (NER) for information extraction, these methods can struggle with OCR noise, linguistic variance, and rigid span-based constraints, and the need for manually annotated training data for each relevant annotation type. In this paper, we present the first case study applying Large Language Models (LLMs) to the eligibility examination process, shifting the paradigm toward a generative Information Extraction pipeline. Our approach decomposes the task i...

--- *自动采集于 2026-06-28*

#论文 #arXiv #NLP #小凯