ChronoFact: Timeline-based Temporal Fact Verification
Anab Maulana Barik, Wynne Hsu, Mong Li Lee
TL;DR
ChronoFact tackles the challenge of verifying complex temporal claims by constructing and aligning chronological timelines for claim and evidence. The approach combines a multi-level attention encoder with event-level, token-level, and time-level analyses, guided by a dual loss framework that enforces per-event, chronological, and overall claim consistency via soft logic using the Gödel $t$-norm. A new ChronoClaims dataset, derived from Wikidata with explicit and implicit temporal cues, enables robust evaluation of timeline-based reasoning, including overlapping and recurring events. Empirical results show ChronoFact achieving state-of-the-art macro F1 on ChronoClaims and strong performance on other temporal benchmarks, with ablations illustrating the importance of chronology modeling. The work advances temporal fact verification by providing a dedicated dataset and a scalable, timeline-aware framework suited to real-world misinformation scenarios.
Abstract
Temporal claims, often riddled with inaccuracies, are a significant challenge in the digital misinformation landscape. Fact-checking systems that can accurately verify such claims are crucial for combating misinformation. Current systems struggle with the complexities of evaluating the accuracy of these claims, especially when they include multiple, overlapping, or recurring events. We introduce a novel timeline-based fact verification framework that identify events from both claim and evidence and organize them into their respective chronological timelines. The framework systematically examines the relationships between the events in both claim and evidence to predict the veracity of each claim event and their chronological accuracy. This allows us to accurately determine the overall veracity of the claim. We also introduce a new dataset of complex temporal claims involving timeline-based reasoning for the training and evaluation of our proposed framework. Experimental results demonstrate the effectiveness of our approach in handling the intricacies of temporal claim verification.
