Code Clone Detection using Machine Learning: Brief Overview and Latest Developments

As software engineering progresses and the demand for code increases, code clones have become more prevalent. Vulnerability propagation is one of the risks posed by this phenomenon, which highlights the increasing significance of code clone detection techniques. It is a highly prevalent method of source code reuse and duplication in source code development. When a bug is detected in a specific snippet of code, it becomes necessary to examine all similar snippets for the presence of the same bug. As a result, this replication procedure may result in the propagation of bugs, which has a substantial impact on maintenance expenses. Owing to this, Code Clone Detection (CCD) emerges as a dynamic field of study. As a result, there is an urgent requirement to examine the most recent methodologies and developments within the field of CCD. In this paper we exhaustively examine the most recent techniques employed in the identification of clones in source code using Machine Learning. The field of Machine learning (ML) has become a preferred approach in the field Artificial Intelligence (AI), and specifically Deep Learning in the field ML, for the development of practical software in numerous fields; detection of source code clones as in our case of study.