AI-Based Program Slicing: Techniques, Tools, and Comparative Analysis

Program slicing is a key software engineering technique that facilitates debugging, maintenance, testing, and security analysis through partitioning slices of relevant code according to slicing criteria. Classic methods of slicing are based on static analysis and dynamic analysis techniques, which might be computationally expensive and lack scalability. Emerging developments in artificial intelligence (AI) have spawned novel program slicing methods based on machine learning (ML), deep learning (DL), natural language processing (NLP), and reinforcement learning (RL). The present paper reviews AI-powered program slicing tools and methods based on performance, pros, and cons. Furthermore, a comparison of AI-enabled program slicing tools and the primary performance measures, including accuracy, efficiency, scalability, and interpretability, is discussed. Lastly, we present existing challenges and future directions in AI-program slicing.