Wide and Deep Reinforcement Learning for Grid-based Action Games.
Juan M. MontoyaChristian BorgeltPublished in: ICAART (2) (2019)
Keyphrases
- reinforcement learning
- action sets
- action selection
- extensive form games
- learning agents
- reward shaping
- partially observable domains
- markov decision processes
- function approximation
- state action
- action space
- machine learning
- reinforcement learning algorithms
- fitted q iteration
- nash equilibria
- temporal difference
- learning algorithm
- state space
- model free
- wide range
- learning process
- game playing
- multi agent
- computer games
- game theory
- optimal policy
- reinforcement learning agents
- transition model
- stochastic games
- neural network
- agent receives
- continuous state
- dynamic environments
- nash equilibrium
- imperfect information
- digital games
- video games
- human actions
- game play
- optimal control