Uncovering Inconsistencies and Contradictions in Financial Reports using Large Language Models
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.
- Veröffentlicht in:
2023 IEEE International Conference on Big Data (BigData) - Typ:
Inproceedings - Autoren:
- Jahr:
2023 - Source:
https://ieeexplore.ieee.org/document/10386673
Informationen zur Zitierung
: Uncovering Inconsistencies and Contradictions in Financial Reports using Large Language Models, 2023 IEEE International Conference on Big Data (BigData), 2023, 2814--2822, https://ieeexplore.ieee.org/document/10386673, Deusser.etal.2023c,
@Inproceedings{Deusser.etal.2023c,
author={Deußer, Tobias; Leonhard, David; Hillebrand, Lars; Berger, Armin; Khaled, Mohamed; Heiden, Sarah; Dilmaghani, Tim; Kliem, Bernd; Loitz, Rüdiger; Bauckhage, Christian; Sifa, Rafet},
title={Uncovering Inconsistencies and Contradictions in Financial Reports using Large Language Models},
booktitle={2023 IEEE International Conference on Big Data (BigData)},
pages={2814--2822},
url={https://ieeexplore.ieee.org/document/10386673},
year={2023},
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...}}
