ViWikiFC: Fact-Checking for Vietnamese Wikipedia-Based Textual Knowledge Source

TL;DR

ViWikiFC fills a critical gap in Vietnamese NLP by introducing the first large-scale, manually annotated Wikipedia-based fact-checking corpus for Vietnamese. The authors construct 20,916 claims across 73 Wikipedia articles, label them with SUPPORTS, REFUTES, or NEI, and provide rigorous annotator and author validation alongside extensive corpus analyses. Empirical evaluation shows BM25 and InfoXLM Large offer the strongest end-to-end performance, though strict accuracy remains around 67%, underscoring the dataset's challenge and the need for improved Vietnamese fact-checking models. This work lays a valuable foundation for future research in low-resource fact-checking, with potential extensions to multi-modal data and broader NLP tasks.

Abstract

Fact-checking is essential due to the explosion of misinformation in the media ecosystem. Although false information exists in every language and country, most research to solve the problem mainly concentrated on huge communities like English and Chinese. Low-resource languages like Vietnamese are necessary to explore corpora and models for fact verification. To bridge this gap, we construct ViWikiFC, the first manual annotated open-domain corpus for Vietnamese Wikipedia Fact Checking more than 20K claims generated by converting evidence sentences extracted from Wikipedia articles. We analyze our corpus through many linguistic aspects, from the new dependency rate, the new n-gram rate, and the new word rate. We conducted various experiments for Vietnamese fact-checking, including evidence retrieval and verdict prediction. BM25 and InfoXLM (Large) achieved the best results in two tasks, with BM25 achieving an accuracy of 88.30% for SUPPORTS, 86.93% for REFUTES, and only 56.67% for the NEI label in the evidence retrieval task, InfoXLM (Large) achieved an F1 score of 86.51%. Furthermore, we also conducted a pipeline approach, which only achieved a strict accuracy of 67.00% when using InfoXLM (Large) and BM25. These results demonstrate that our dataset is challenging for the Vietnamese language model in fact-checking tasks.

Figures (12)

• Figure 1: The creation process of the fact-checking corpus.
• Figure 2: ViWifiFC annotation tool for fact-checking corpus.
• Figure 3: ViWifiFC validation tool for fact-checking corpus.
• Figure 4: The distribution total syllable in the corpus.
• Figure 5: The distribution total word in the corpus.