A Convolutional Attention Network for Extreme Summarization of Source Code

Literature review

This is a paper mentioned in the course on multilingual NLP by Graham Neubig. With an interesting idea of an second attention head being used to copy stuff from the input directly to the output.

Abstract

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. – (Allamanis, Peng, and Sutton 2016)

Outline

The Paper

paper

References

Allamanis, Miltiadis, Hao Peng, and Charles Sutton. 2016. “A Convolutional Attention Network for Extreme Summarization of Source Code.” https://arxiv.org/abs/1602.03001.