Code clone detection framework based on domain adaptation

Abstract

Code clones are duplicate code fragments that share (nearly) similar syntax or semantics. Code clone detection plays an important role in software maintenance, code refactoring, and reuse. Most of the techniques for code clone detection have achieved unprecedented performance on different open-source data sets. However, these methods are predominantly supervised and work only on samples drawn from the same distribution on which they have been trained. Since there is a scarcity of labeled data of code clones, it becomes hard to use these techniques in real-world software systems. To overcome this limitation, this project plans to build a code clone detection framework based on domain adaptation. I am planning to build this on top of an existing code clone detection tool - HOLMES.