Topic-Aware Evidence Reasoning and Stance-Aware Aggregation for Fact Verification

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 1612–1622 August 1–6, 2021. ©2021 Association for Computational Linguistics 1612 Topic-Aware Evidence Reasoning and Stance-Aware Aggregation for Fact Verification Jiasheng Si† Deyu Zhou†∗ Tongzhe Li† Xingyu Shi† Yulan He§ † School of Computer Science and Engineering, Key Laboratory of Computer Network and Information Integration, Ministry of Education, Southeast University, China § Department of Computer Science, University of Warwick, UK {jasenchn, d.zhou, 220184611, xyu-shi}@seu.edu.cn, yulan.he@warwick.ac.uk Abstract Fact verification is a challenging task that re- quires simultaneously reasoning and aggregat- ing over multiple retrieved pieces of evidence to evaluate the truthfulness of a claim. Ex- isting approaches typically (i) explore the se- mantic interaction between the claim and ev- idence at different granularity levels but fail to capture their topical consistency during the reasoning process, which we believe is cru- cial for verification; (ii) aggregate multiple pieces of evidence equally without considering their implicit stances to the claim, thereby in- troducing spurious information. To alleviate the above issues, we propose a novel topic- aware evidence reasoning and stance-aware ag- gregation model for more accurate fact veri- fication, with the following four key proper- ties: 1) checking topical consistency between the claim and evidence; 2) maintaining topical coherence among multiple pieces of evidence; 3) ensuring semantic similarity between the global topic information and the semantic rep- resentation of evidence; 4) aggregating evi- dence based on their implicit stances to the claim. Extensive experiments conducted on the two benchmark datasets demonstrate the superiority of the proposed model over sev- eral state-of-the-art approaches for fact verifi- cation. The source code can be obtained from https://github.com/jasenchn/TARSA. 1 Introduction The Internet breaks the physical distance barrier among individuals to allow them to share data and information online. However, it can also be used by people with malicious purposes to disseminate misinformation or fake news. Such misinformation may cause ethnics conflicts, financial losses and po- litical unrest, which has become one of the greatest threats to the public (Zafarani et al., 2019; Zhou ∗corresponding author et al., 2019b). Moreover, as shown in Vosoughi et al. (2018), compared with truth, misinformation diffuses significantly farther, faster, and deeper in all genres. Therefore, there is an urgent need for quickly identifying the misinformation spread on the web. To solve this problem, we focus on the fact verification task (Thorne et al., 2018), which aims to automatically evaluate the veracity of a given claim based on the textual evidence retrieved from external sources. Recent approaches for fact verification are domi- nated by natural language inference models (Angeli and Manning, 2014) or textual entailment recogni- tion models (Ma et al., 2019), where the truthful- ness of a claim is verified via reasoning and aggre- gating over multiple pieces of retrieved evidence. In general, existing models follow an architecture with two main sub-modules: the semantic inter- action module and the entailment-based aggrega- tion module (Hanselowski et al., 2018a; Nie et al., …

Content truncated to 500 words