A Transformer-based Approach for Source Code Summarization

TL;DR

This paper presents a Transformer-based method for source code summarization that effectively captures long-range dependencies, outperforming previous techniques and emphasizing the importance of relative positional encoding.

Contribution

The study demonstrates that a simple Transformer model with relative positional encoding significantly improves source code summarization performance over existing methods.

Findings

Relative positional encoding enhances summarization accuracy.

Transformer model outperforms state-of-the-art techniques.

Code is publicly available for future research.

Abstract

Generating a readable summary that describes the functionality of a program is known as source code summarization. In this task, learning code representation by modeling the pairwise relationship between code tokens to capture their long-range dependencies is crucial. To learn code representation for summarization, we explore the Transformer model that uses a self-attention mechanism and has shown to be effective in capturing long-range dependencies. In this work, we show that despite the approach is simple, it outperforms the state-of-the-art techniques by a significant margin. We perform extensive analysis and ablation studies that reveal several important findings, e.g., the absolute encoding of source code tokens' position hinders, while relative encoding significantly improves the summarization performance. We have made our code publicly available to facilitate future research.