COVID-VTS: Fact Extraction and Verification on Short Video Platforms

TL;DR

This paper presents COVID-VTS, a benchmark for fact-checking COVID-19 related short videos, and introduces TwtrDetective, a model that detects tampering and verifies multi-modal information with explanations.

Contribution

The paper introduces COVID-VTS benchmark and TwtrDetective model, advancing multi-modal fact verification and tampering detection in short videos with limited training data.

Findings

TwtrDetective outperforms state-of-the-art models

Efficient automatic generation of misleading videos

Improved multi-modal tampering detection

Abstract

We introduce a new benchmark, COVID-VTS, for fact-checking multi-modal information involving short-duration videos with COVID19- focused information from both the real world and machine generation. We propose, TwtrDetective, an effective model incorporating cross-media consistency checking to detect token-level malicious tampering in different modalities, and generate explanations. Due to the scarcity of training data, we also develop an efficient and scalable approach to automatically generate misleading video posts by event manipulation or adversarial matching. We investigate several state-of-the-art models and demonstrate the superiority of TwtrDetective.