Towards Human-Like RL: Taming Non-Naturalistic Behavior in Deep RL via Adaptive Behavioral Costs in 3D Games.
Kuo-Hao HoPing-Chun HsiehChiu-Chou LinYou-Ren LuoFeng-Jian WangI-Chen WuPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- human behavior
- learning agents
- model free
- adaptive control
- adaptive behavior
- reinforcement learning algorithms
- learning classifier systems
- markov decision processes
- autonomous learning
- behavioral model
- partially observable domains
- state space
- reinforcement learning agents
- optimal policy
- learning algorithm
- action space
- dynamic programming
- behavior patterns
- multi agent
- actor critic
- behavioral patterns
- total cost
- transfer learning