A Literature Review on Automatic Comment Generation using Deep Learning Techniques

Abstract

During the development of software, thousands of lines of codes are written and without proper comments it gets difficult for the developers to read the code. Automatic Comment Generation overcomes this problem by generating comments for the source code. Many researchers are already working on automatic comment generation in the source code and some have already published their results but still there is lot of scope of improvement in the field of automatic generation of comments in the source code. In this paper, we will be explaining and reviewing deep learning techniques that areused in comment generation to make the code easily readable. Generating comments automatically become crucial in the source code for any developer to understand it easily without spending much time. Hence, it's very important to understand how deep learning techniques like CNN, RNN, LSTM and others are making it possible to generate automatic comments in source code and also why researchers have now shifted on deep learning techniques to do automatic comment generation.