BERT Models for Arabic Text Classification: A Systematic Review

Bidirectional Encoder Representations from Transformers (BERT) has gained increasing attention researchers and practitioners as it proven to be an invaluable technique in natural languages processing. This is mainly due its unique features, including ability predict words conditioned on both the left right context, pretrained using plain text corpus that enormously available web. As BERT more interest, models were introduced support different languages, Arabic. The current state of knowledge practice applying Arabic classification limited. In attempt begin remedying this gap, review synthesizes have been applied classification. It investigates differences between them compares their performance. also examines how effective they are compared original English models. concludes by offering insight into aspects need further improvements future work.