Semantic Textual Similarity: Overview and Comparative Study between Arabic and English

Abstract

Semantic Textual Similarity is crucial for various end-user applications of Natural Language Processing, including Search Engines, Chatbots, Machine Translation Systems, Plagiarism Detection, and Text Summarization. While substantial research has been conducted on this topic for widely spoken languages such as English, there exists a need for comprehensive surveys focusing on less-studied languages, such as Arabic. This work is a comprehensive resource for researchers working on Semantic Textual Similarity especially for the Arabic language.Our survey synthesizes the current state of research in Semantic Textual Similarity in Arabic, providing valuable insights into this field's unique challenges and opportunities. We review state-of-the-art approaches, datasets, and methodologies proposed for Arabic Semantic Textual Similarity. The paper highlights the differences between Arabic and English, which necessitate tailored approaches to Semantic Textual Similarity. Moreover, we discuss the recent advancements in Arabic Semantic Textual Similarity and identify the existing gaps and challenges that researchers face. In addition, we propose potential future research directions to further improve the Arabic Semantic Textual Similarity models. By addressing these areas, our work aims to foster a deeper understanding and more robust development of Semantic Textual Similarity for the Arabic language, ultimately expanding the scope and effectiveness of Semantic Textual Similarity applications.