From Video Game to Real Robot: The Transfer Between Action Spaces.
Janne KarttunenAnssi KanervistoVille KyrkiVille HautamäkiPublished in: ICASSP (2020)
Keyphrases
- video games
- real robot
- action space
- reinforcement learning
- state space
- mobile robot
- skill learning
- robot arm
- state and action spaces
- markov decision processes
- learning experience
- simulated robot
- real valued
- educational games
- stochastic processes
- single agent
- action selection
- game play
- multi robot
- motion control
- continuous action
- multi agent
- learning process
- real world
- robotic systems
- computer games
- humanoid robot
- markov decision process
- optimal policy
- game design
- multiple agents
- dynamic programming
- control system
- control law
- serious games
- game players
- learning algorithm
- data sets
- machine learning
- image sequences
- real time