Exploring Decomposition for Table-based Fact Verification

Xiaoyu Yang, Xiaodan Zhu


Abstract
Fact verification based on structured data is challenging as it requires models to understand both natural language and symbolic operations performed over tables. Although pre-trained language models have demonstrated a strong capability in verifying simple statements, they struggle with complex statements that involve multiple operations. In this paper, we improve fact verification by decomposing complex statements into simpler subproblems. Leveraging the programs synthesized by a weakly supervised semantic parser, we propose a program-guided approach to constructing a pseudo dataset for decomposition model training. The subproblems, together with their predicted answers, serve as the intermediate evidence to enhance our fact verification model. Experiments show that our proposed approach achieves the new state-of-the-art performance, an 82.7% accuracy, on the TabFact benchmark.
Anthology ID:
2021.findings-emnlp.90
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2021
Month:
November
Year:
2021
Address:
Punta Cana, Dominican Republic
Venues:
EMNLP | Findings
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
1045–1052
Language:
URL:
https://aclanthology.org/2021.findings-emnlp.90
DOI:
10.18653/v1/2021.findings-emnlp.90
Bibkey:
Cite (ACL):
Xiaoyu Yang and Xiaodan Zhu. 2021. Exploring Decomposition for Table-based Fact Verification. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1045–1052, Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Exploring Decomposition for Table-based Fact Verification (Yang & Zhu, Findings 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.findings-emnlp.90.pdf
Code
 arielsho/decomposition-table-reasoning
Data
TabFact