Combat COVID-19 infodemic using explainable natural language processing models

• Published in: Information processing & management Vol. 58; no. 4; p. 102569
• Main Authors: Ayoub, Jackie, Yang, X. Jessie, Zhou, Feng
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.07.2021; Elsevier Science Ltd
• Subjects: BERT; Between-subjects design; Coders; Coronaviruses; COVID-19; Datasets; Deep learning; DistilBERT; False information; Machine learning; Misinformation detection; Model testing; Natural language processing; SHAP; Social networks; Trust; COVID-19; SHAP; BERT; Misinformation detection; DistilBERT; Trust
• ISSN: 0306-4573; 1873-5371; 0306-4573
• DOI: 10.1016/j.ipm.2021.102569

Misinformation of COVID-19 is prevalent on social media as the pandemic unfolds, and the associated risks are extremely high. Thus, it is critical to detect and combat such misinformation. Recently, deep learning models using natural language processing techniques, such as BERT (Bidirectional Encoder Representations from Transformers), have achieved great successes in detecting misinformation. In this paper, we proposed an explainable natural language processing model based on DistilBERT and SHAP (Shapley Additive exPlanations) to combat misinformation about COVID-19 due to their efficiency and effectiveness. First, we collected a dataset of 984 claims about COVID-19 with fact-checking. By augmenting the data using back-translation, we doubled the sample size of the dataset and the DistilBERT model was able to obtain good performance (accuracy: 0.972; areas under the curve: 0.993) in detecting misinformation about COVID-19. Our model was also tested on a larger dataset for AAAI2021 — COVID-19 Fake News Detection Shared Task and obtained good performance (accuracy: 0.938; areas under the curve: 0.985). The performance on both datasets was better than traditional machine learning models. Second, in order to boost public trust in model prediction, we employed SHAP to improve model explainability, which was further evaluated using a between-subjects experiment with three conditions, i.e., text (T), text+SHAP explanation (TSE), and text+SHAP explanation+source and evidence (TSESE). The participants were significantly more likely to trust and share information related to COVID-19 in the TSE and TSESE conditions than in the T condition. Our results provided good implications for detecting misinformation about COVID-19 and improving public trust. •Natural language processing models based on DistilBERT are able to detect misinformation of COVID-19.•Explanations based on SHAP are able to improve user trust in model prediction.•With extra explanations on claims of COVID-19, users are more willing to share them.•It is important not only to detect misinformation but also to explain why it is fake or true.