A Convolutional Attention Network for Extreme Summarization of Source Code
This addresses the challenge of extreme summarization for developers, but it is incremental as it builds on existing attentional architectures with a novel hybrid approach.
The paper tackles the problem of generating short, descriptive summaries for source code snippets by introducing a convolutional attention network that uses convolution to detect local and long-range attention features, achieving better performance than previous attentional mechanisms on 10 popular Java projects.
Attention mechanisms in neural networks have proved useful for problems in which the input and output do not have fixed dimension. Often there exist features that are locally translation invariant and would be valuable for directing the model's attention, but previous attentional architectures are not constructed to learn such features specifically. We introduce an attentional neural network that employs convolution on the input tokens to detect local time-invariant and long-range topical attention features in a context-dependent way. We apply this architecture to the problem of extreme summarization of source code snippets into short, descriptive function name-like summaries. Using those features, the model sequentially generates a summary by marginalizing over two attention mechanisms: one that predicts the next summary token based on the attention weights of the input tokens and another that is able to copy a code token as-is directly into the summary. We demonstrate our convolutional attention neural network's performance on 10 popular Java projects showing that it achieves better performance compared to previous attentional mechanisms.