Abstract
The rapid advances of natural language processing (NLP) research bring about applications and tools that serve automating Arabic at different linguistic levels. However, there is a focus on the modern standard Arabic (MSA) more than the dialects which are used heavily in day-to-day communication and require special handling. This paper introduces research efforts in Arabic NLP with emphasis on tools, applications, and resources related to core NLP areas. Some research and scholarly work come from researchers working on extended projects to compute Arabic. The findings demonstrated that research on Arabic NLP can be characterized by emphasis on morphological analysis and processing of MSA especially tokenization and tagging. Recently there has been a growing interest in processing and identifying Arabic dialects. The proposed solutions were adapted from other solutions used for MSA or designed mainly for certain dialects. Also, advances in developing transformers were applied to Arabic NLP in limited research. It appeared that ambiguity is still a big challenge to overcome and there is a need to innovate applications and tools for Arabic MSA and the dialects in parallel.