Optimizing Arabic Named Entity Recognition through Active Learning and AraBERT

This study explores advancements in Arabic named entity recognition (NER) through the application of active learning techniques and the utilization of large language models, including AraBERT. The research investigates the effectiveness of active learning in selecting informative instances for annotation while leveraging the power of AraBERT to enhance NER performance. The study aims to adapt AraBERT to handle the unique characteristics and requirements of Arabic text. By employing these approaches, the research contributes to the development of more accurate NER systems for Arabic, addressing challenges related to limited labeled data.