Code Recommendation for Open Source Project Developers
Abstract
Open Source Software (OSS) developers contribute to a large and rapidly growing ecosystem, but matching developers to relevant code is challenging. We introduce CODER, a graph-based code recommendation framework that leverages multi-modal signals from OSS contribution and engagement networks. CODER models heterogeneous graphs over repositories, users, issues, and pull requests, and learns representations that improve code recommendation across realistic OSS workloads. WWW 2023 acceptance rate: 19.2%.
Type
Publication
The Web Conference (WWW) 2023, Oral Presentation
Abstract
Open Source Software (OSS) developers contribute to a large and rapidly growing ecosystem, but matching developers to relevant code is challenging. We introduce CODER, a graph-based code recommendation framework that leverages multi-modal signals from OSS contribution and engagement networks.
