Discrete Action On-Policy Learning with Action-Value Critic.
Yuguang YueYunhao TangMingzhang YinMingyuan ZhouPublished in: AISTATS (2020)
Keyphrases
- action selection
- learning process
- learning algorithm
- reinforcement learning
- action models
- learning problems
- state action
- actor critic
- partially observable domains
- learning systems
- learning from experience
- knowledge acquisition
- learning experience
- machine learning
- supervised learning
- mobile robot
- prior knowledge
- policy gradient
- continuous state spaces
- inverse reinforcement learning