An Empirical Study on Learning-based Techniques for Explicit and Implicit Commit Messages Generation
Zhiquan Huang, Yuan Huang, Xiangping Chen, Xiaocong Zhou, Changlin Yang, Zibin Zheng · 2024
High-quality and appropriate commit messages help developers to quickly understand and track code evolution, which is crucial for the collaborative development and maintenance of software. To relieve developers of the burden of writing commit messages, researchers have proposed various techniques to generate commit messages automatically, among which learning-based techniques have proven to be promising.