Code-to-Code Search Based on Deep Neural Network and Code Mutation

Deep Neural Networks (DNNs) have been often used for the labeling of image files (e.g., object detection). Although they can be applied for the labeling of code fragment (i.e., code-to-code search) in software engineering, a large number of code fragments are required for each label in the learning process of DNNs. In this paper, we present an approach for code-to-code search based on a DNN model and code mutation for generating enough number of code fragments for each label. The preliminary experiment shows high precision and recall of the proposed approach.