Zero-shot Faithful Factual Error Correction

Residential College	false
Status	已發表Published
	Zero-shot Faithful Factual Error Correction
	Huang, Kung Hsiang 1; Chan, Hou Pong 2; Ji, Heng 1
	2023
Source Publication	Proceedings of the Annual Meeting of the Association for Computational Linguistics
Volume	1
Pages	5660-5676
Abstract	Faithfully correcting factual errors is critical for maintaining the integrity of textual knowledge bases and preventing hallucinations in generative models. Drawing on humans' ability to identify and correct factual errors, we present a zero-shot framework that formulates questions about input claims, looks for correct answers in the given evidence, and assesses the faithfulness of each correction based on its consistency with the evidence. Our zero-shot framework outperforms fully-supervised approaches, as demonstrated by experiments on the FEVER and SCIFACT datasets, where our outputs are shown to be more faithful. More importantly, the decomposability nature of our framework inherently provides interpretability. Additionally, to reveal the most suitable metrics for evaluating factual error corrections, we analyze the correlation between commonly used metrics with human judgments in terms of three different dimensions regarding intelligibility and faithfulness.
URL	View the original
Language	英語English
Scopus ID	2-s2.0-85172430751
Fulltext Access	View Full-Text via Scopus
Citation statistics
Document Type	Conference paper
Collection	University of Macau
Affiliation	1.Department of Computer Science, University of Illinois Urbana-Champaign, United States 2.Faculty of Science and Technology, University of Macau, Macao
Recommended Citation GB/T 7714	Huang, Kung Hsiang,Chan, Hou Pong,Ji, Heng. Zero-shot Faithful Factual Error Correction[C], 2023, 5660-5676.
APA	Huang, Kung Hsiang., Chan, Hou Pong., & Ji, Heng (2023). Zero-shot Faithful Factual Error Correction. Proceedings of the Annual Meeting of the Association for Computational Linguistics, 1, 5660-5676.

Files in This Item:
There are no files associated with this item.

If you have any objections to this item, please fill out the form below and the administrator will contact you as soon as possible.
Content:
Email：	*
Affiliation No.
Verification Code:	Refresh

Any comments and suggestions are welcomed.
Title:	*
Content:
Email：	*
Verification Code:	Refresh