Review Paper on Enhancing COVID-19 Fake News Detection With Transformer Model

Abstract

The growing propagation of disinformation about the COVID-19 epidemic needs powerful fake news detection technologies. This review provides an in-depth examination of existing techniques, including traditional machine learning methods such as Random Forest and Naive Bayes, as well as sophisticated models for deep learning such as Bi- GRU, CNN, and LSTM, RNN, & transformer-based architecture such as BERT and XLM- Roberta, are also available. One noticeable development is the merging of traditional algorithmswith sophisticated transformers, which emphasize the quest of improved accuracy and flexibility.However, important research gaps have been identified. There has been little research on cross- lingual detection algorithms, revealing a substantial gap in multilingual false news detection, which is critical in the global context of COVID-19 information spread. Furthermore, the researchemphasizes the need of flexible methodologies by emphasizing the need for appropriate preprocessing strategies for various content types. Furthermore, the lack of common assessment measures is a barrier, underlining the need of unified frameworks for successfully benchmarking and comparing models. This analysis provides light on the changing COVID-19 false news detection environment, emphasizing the need for novel, adaptive, and internationally relevant approaches to successfully address the ubiquitous dissemination of disinformation during the current pandemic.