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

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.

References in corpus (5)

• Efficient Estimation of Word Representations in Vector Space
• Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads
• Discourse-Aware Rumour Stance Classification in Social Media Using Sequential Classifiers
• A Multi-task Approach for Named Entity Recognition in Social Media Data
• Determining the Veracity of Rumours on Twitter

Cited by in corpus (23)

• Combating Fake News: A Survey on Identification and Mitigation Techniques
• Integrating Pattern- and Fact-based Fake News Detection via Model Preference Learning
• Deep Two-path Semi-supervised Learning for Fake News Detection
• Multi-Task Learning in Natural Language Processing: An Overview
• Dataset of Fake News Detection and Fact Verification: A Survey
• Adaptive Interaction Fusion Networks for Fake News Detection
• Neural Language Model Based Training Data Augmentation for Weakly Supervised Early Rumor Detection
• A Survey on Predicting the Factuality and the Bias of News Media
• Let-Mi: An Arabic Levantine Twitter Dataset for Misogynistic Language
• DTCA: Decision Tree-based Co-Attention Networks for Explainable Claim Verification
• QMUL-SDS at CheckThat! 2020: Determining COVID-19 Tweet Check-Worthiness Using an Enhanced CT-BERT with Numeric Expressions
• Two-path Deep Semi-supervised Learning for Timely Fake News Detection
• MALCOM: Generating Malicious Comments to Attack Neural Fake News Detection Models
• Towards Propagation Uncertainty: Edge-enhanced Bayesian Graph Convolutional Networks for Rumor Detection
• Dynamically Addressing Unseen Rumor via Continual Learning
• Meet The Truth: Leverage Objective Facts and Subjective Views for Interpretable Rumor Detection
• Modeling Conversation Structure and Temporal Dynamics for Jointly Predicting Rumor Stance and Veracity
• Contrastive Language Adaptation for Cross-Lingual Stance Detection
• Rumour Detection via Zero-shot Cross-lingual Transfer Learning
• Towards Explainable Fact Checking
• Detecting Toxicity in News Articles: Application to Bulgarian
• Leveraging Commonsense Knowledge on Classifying False News and Determining Checkworthiness of Claims
• Measuring What Counts: The case of Rumour Stance Classification