Open-Source AI-based SE Tools: Opportunities and Challenges of Collaborative Software Learning

• DOI: 10.48550/arxiv.2404.06201
• Publication Date: 2024-04-09
• Authors: Zhihao Lin; Wei Ma; Tao Lin; Yaowen Zheng; Jingquan Ge; Jun Wang; Jacques Klein; Tegawendé F. Biss...; Liu Yang; Li Li
• Keywords: Software Engineering (cs.SE); FOS: Computer and information sciences; Computer Science - Software Engineering; Artificial Intelligence (cs.AI); Computer Science - Artificial Intelligence; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

Large Language Models (LLMs) have become instrumental in advancing software engineering (SE) tasks, showcasing their efficacy code understanding and beyond. Like traditional SE tools, open-source collaboration is key realising the excellent products. However, with AI models, essential need data. The of these AI-based models hinges on maximising sources high-quality data especially high quality, often holds commercial or sensitive value, making it less accessible for projects. This reality presents a significant barrier to development enhancement tools within community. Therefore, researchers find solutions enabling tap into resources by different organisations. Addressing this challenge, our position paper investigates one solution facilitate access diverse organizational ensuring privacy sensitivities are respected. We introduce governance framework centered federated learning (FL), designed foster joint maintenance while safeguarding security. Additionally, we present guidelines developers tool collaboration, covering requirements, model architecture, updating strategies, version control. Given influence characteristics FL, research examines effect heterogeneity FL performance.