"Beware of deception": Detecting Half-Truth and Debunking it through Controlled Claim Editing

TL;DR

This paper introduces a pipeline combining half-truth detection and claim editing using a T5 model, significantly improving detection accuracy and enabling effective debunking of deceptive statements on the internet.

Contribution

It presents the first integrated approach for detecting and debunking half-truths, utilizing controlled claim editing with a T5 model and extending datasets for better performance.

Findings

Achieved an F1 score of 82% in half-truth detection.

T5-based claim editing outperforms GPT2, RoBERTa, PEGASUS, and Tailor.

Edited claims received an 85% disinfo-debunk score.

Abstract

The prevalence of half-truths, which are statements containing some truth but that are ultimately deceptive, has risen with the increasing use of the internet. To help combat this problem, we have created a comprehensive pipeline consisting of a half-truth detection model and a claim editing model. Our approach utilizes the T5 model for controlled claim editing; "controlled" here means precise adjustments to select parts of a claim. Our methodology achieves an average BLEU score of 0.88 (on a scale of 0-1) and a disinfo-debunk score of 85% on edited claims. Significantly, our T5-based approach outperforms other Language Models such as GPT2, RoBERTa, PEGASUS, and Tailor, with average improvements of 82%, 57%, 42%, and 23% in disinfo-debunk scores, respectively. By extending the LIAR PLUS dataset, we achieve an F1 score of 82% for the half-truth detection model, setting a new benchmark in the field. While previous attempts have been made at half-truth detection, our approach is, to the best of our knowledge, the first to attempt to debunk half-truths.