Automatic Generation of Text Descriptive Comments for Code Blocks
This work addresses the challenge of improving code documentation for developers, but it is incremental as it builds on existing learning-based methods with specific enhancements.
The authors tackled the problem of automatically generating descriptive comments for source code blocks by proposing a framework that uses a recursive neural network (Code-RNN) to extract features and a recurrent neural network (Code-GRU) to generate text, achieving significantly higher accuracy with a Rouge-2 value compared to other learning-based approaches.
We propose a framework to automatically generate descriptive comments for source code blocks. While this problem has been studied by many researchers previously, their methods are mostly based on fixed template and achieves poor results. Our framework does not rely on any template, but makes use of a new recursive neural network called Code-RNN to extract features from the source code and embed them into one vector. When this vector representation is input to a new recurrent neural network (Code-GRU), the overall framework generates text descriptions of the code with accuracy (Rouge-2 value) significantly higher than other learning-based approaches such as sequence-to-sequence model. The Code-RNN model can also be used in other scenario where the representation of code is required.