Uncovering Inconsistencies and Contradictions in Financial Reports using Large Language Models
Authors:
- 15 December 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2814-2822
Abstract
Correct identification and correction of contradictions and inconsistencies within financial reports constitute a fundamental component of the audit process. To streamline and automate this critical task, we introduce a novel approach leveraging large language models and an embedding-based paragraph clustering methodology. This paper assesses our approach across three distinct datasets, including two annotated datasets and one unannotated dataset, all within a zero-shot framework. Our findings reveal highly promising results that significantly enhance the effectiveness and efficiency of the auditing process, ultimately reducing the time required for a thorough and reliable financial report audit.Keywords
Funding Information
- Ministry of Education
This publication has 32 references indexed in Scilit:
- “Here Are the Rules: Ignore All Rules”: Automatic Contradiction Detection in SpanishApplied Sciences, 2021
- Tackling Contradiction Detection in German Using Machine Translation and End-to-End Recurrent Neural Networks2021
- SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization2020
- Towards Contradiction Detection in German: a Translation-Driven Approach2019
- Towards Automated Auditing with Machine Learning2019
- Towards Automatic Numerical Cross-Checking2018
- Deep Gaussian mixture modelsStatistics and Computing, 2017
- Classification and Acquisition of Contradictory Event Pairs using Crowdsourcing2015
- Density-Based Clustering Based on Hierarchical Density Estimates2013
- The PASCAL Recognising Textual Entailment Challenge2006