Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey of Code-switched Arabic NLP: Progress, Challenges, and Future Directions

Injy Hamed, Caroline Sabty, Slim Abdennadher, Ngoc Thang Vu, Thamar Solorio, Nizar Habash

TL;DR

This survey analyzes code-switched Arabic NLP within a diglossic and multilingual setting, surveying literature, data resources, tasks, and benchmarks to map progress and gaps. It documents progress across LID, MT, ASR, transliteration, and NLP guidelines, while highlighting late adoption of neural methods and limited benchmarks. Key contributions include a taxonomy of CSW types, a structured annotation framework for language pairs, and a prioritized agenda of future directions (benchmarks, pretrained models, user-facing apps, personalization, and ethics). The findings underscore the need for broader datasets, improved evaluation metrics, and responsible deployment to support Arabic-speaking multilingual communities in practical NLP applications.

Abstract

Language in the Arab world presents a complex diglossic and multilingual setting, involving the use of Modern Standard Arabic, various dialects and sub-dialects, as well as multiple European languages. This diverse linguistic landscape has given rise to code-switching, both within Arabic varieties and between Arabic and foreign languages. The widespread occurrence of code-switching across the region makes it vital to address these linguistic needs when developing language technologies. In this paper, we provide a review of the current literature in the field of code-switched Arabic NLP, offering a broad perspective on ongoing efforts, challenges, research gaps, and recommendations for future research directions.

A Survey of Code-switched Arabic NLP: Progress, Challenges, and Future Directions

TL;DR

This survey analyzes code-switched Arabic NLP within a diglossic and multilingual setting, surveying literature, data resources, tasks, and benchmarks to map progress and gaps. It documents progress across LID, MT, ASR, transliteration, and NLP guidelines, while highlighting late adoption of neural methods and limited benchmarks. Key contributions include a taxonomy of CSW types, a structured annotation framework for language pairs, and a prioritized agenda of future directions (benchmarks, pretrained models, user-facing apps, personalization, and ethics). The findings underscore the need for broader datasets, improved evaluation metrics, and responsible deployment to support Arabic-speaking multilingual communities in practical NLP applications.

Abstract

Language in the Arab world presents a complex diglossic and multilingual setting, involving the use of Modern Standard Arabic, various dialects and sub-dialects, as well as multiple European languages. This diverse linguistic landscape has given rise to code-switching, both within Arabic varieties and between Arabic and foreign languages. The widespread occurrence of code-switching across the region makes it vital to address these linguistic needs when developing language technologies. In this paper, we provide a review of the current literature in the field of code-switched Arabic NLP, offering a broad perspective on ongoing efforts, challenges, research gaps, and recommendations for future research directions.
Paper Structure (38 sections, 3 figures, 11 tables)

This paper contains 38 sections, 3 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Language in the Arab World
  3. Types of Code-switching
  4. Paper Categorization Process
  5. Overview on Literature Review
  6. Number of Papers Across the Years
  7. Distribution of Papers Across Venue Types
  8. Evolution in Methodology Methods
  9. Investigated Language Pairs
  10. Insights on Language Resources
  11. Dataset Sources
  12. NLP Tasks' Coverage
  13. Insights on NLP Tasks
  14. Language Identification (LID)
  15. Word-level LID
  16. ...and 23 more sections

Figures (3)

  • Figure 1: Number of Shared Task (ST) and non-Shared Task (non-ST) papers.
  • Figure 2: Distribution of papers based on the methods.
  • Figure 3: The number of papers covering the different CSW language setups. We present the distribution as per the annotation guidelines (left) with further language specification (right). Language codes are provided in Appendix \ref{['sec:appendix_lang_codes']}.