IRCoCo: Immediate Rewards-Guided Deep Reinforcement Learning for Code Completion.
Bolun LiZhihong SunTao HuangHongyu ZhangYao WanGe LiZhi JinChen LyuPublished in: CoRR (2024)
Keyphrases
- reinforcement learning
- markov decision processes
- function approximation
- source code
- reinforcement learning algorithms
- state space
- temporal difference
- model free
- reward shaping
- learning algorithm
- partially observable
- machine learning
- optimal policy
- supervised learning
- learning process
- markov decision process
- learning problems
- transfer learning
- case study
- reinforcement learning methods
- hidden state
- action selection
- deep learning
- error correcting
- robotic control
- multiarmed bandit