Reward Engineering for Software Tasks: A Survey of Reinforcement Learning Approaches

Md Rayhanul Masud, Azmine Toushik Wasi, Salman Rahman +1

Reinforcement learning is increasingly used for code-centric software engineering tasks, including code generation, understanding, repair, testing, and optimization, especially wit…