All-in-one: Multi-task Learning for Rumour Verification

TL;DR

This paper introduces a multi-task learning framework that jointly trains rumour detection, tracking, stance classification, and veracity determination, enhancing overall rumour verification performance.

Contribution

It presents a novel multi-task learning approach that integrates multiple components of rumour verification into a single model, improving accuracy over separate models.

Findings

Multi-task learning improves rumour verification accuracy.

Dataset properties influence multi-task learning outcomes.

Joint training benefits multiple rumour analysis tasks.

Abstract

Automatic resolution of rumours is a challenging task that can be broken down into smaller components that make up a pipeline, including rumour detection, rumour tracking and stance classification, leading to the final outcome of determining the veracity of a rumour. In previous work, these steps in the process of rumour verification have been developed as separate components where the output of one feeds into the next. We propose a multi-task learning approach that allows joint training of the main and auxiliary tasks, improving the performance of rumour verification. We examine the connection between the dataset properties and the outcomes of the multi-task learning models used.