Vi2FC: A Two-Stage Framework for Automatic Fact Verification on Vietnamese Texts

Do, Khoa Nguyen-Anh, Luu, Son T.
The Review of Socionetwork Strategies · Springer Nature · 2026-10-01 · DOI: 10.1007/s12626-026-00217-w

Abstract

Fact-checking is the process of assessing the truthfulness of a claim. With the rapid spreading of information on social media in Vietnam in recent years, Automatic Fact-checking plays an important role in decreasing the impact of misleading and manipulated information on the Internet. In this paper, we introduce Vi2FC - a two-stage framework for constructing an automated Fact-checking system in the Vietnamese language. The Vi2FC consists of two stages: retrieving the relevant evidence sentence for the claim and verifying the claim based on the evidence. Then, we evaluate the Vi2FC framework on three different Vietnamese Fact-checking datasets, including ViWikiFC, DSC, and ViFactCheck. Our Vi2FC framework achieves competitive results with other state-of-the-art and baseline models, which are 85.55% by macro F1 score on ViWikiFC, 69.29% by Accuracy on DSC, and 85.08% by macro F1 score on the ViFactCheck. Additionally, we evaluate the performance of Vi2FC on gold and system evidence to assess its potential for practical application in Vietnamese fact-checking systems.

Read at publisher →

Similar articles