Semantic Text Matching Using Intelligent Methods: A Survey

Abstract

Nowadays, Artificial intelligence (AI) is emerged strongly. AI applications are the future of computing, with numerous services and solutions. Machine learning (ML) and deep learning (DL) are the most important components of AI-emerged tools. Natural language processing (NLP) is that field related to incorporating AI tools for improving machine recognition of human natural languages. Text mining and information retrieval applications have become important in frequent intelligent business services. Semantic text matching (STM) is an important application of NLP and text mining. It produces an understanding of two or more text segments in terms of similarity and distance evaluation. In this survey, AI-based STM approaches that are introduced in the literature are reviewed. We present the importance of the STM field using intelligent methods, in addition to explaining the main STMbased concepts. We discuss various categories of STM that appear in surveyed papers, they are the type of language, the deep learning utilized models, question-answering interesting studies, and other smart techniques that were implemented for STM. Finally, we summarize this review by introducing an explanation of some recent papers, in addition to illustrating a comparison between them. The utilized dataset, implemented model, and resulted accuracy were the main points of that comparison.