Deriving Rewards for Reinforcement Learning from Symbolic Behaviour Descriptions of Bipedal Walking.
Daniel HarnackChristoph LüthLukas GrossShivesh KumarFrank KirchnerPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- reward function
- symbolic description
- high level
- markov decision processes
- function approximation
- state space
- reinforcement learning algorithms
- symbolic descriptions
- policy search
- multi agent
- model free
- partially observable
- temporal difference
- function approximators
- machine learning
- markov decision process
- optimal policy
- partial observability
- connectionist systems
- robotic control
- learning algorithm
- action selection
- user behaviour
- reinforcement learning methods
- supervised learning
- dynamic programming
- complex domains
- symbolic representation
- semantic description
- real robot
- hidden state
- autonomous learning
- connectionist models
- neural learning
- transfer learning
- natural language descriptions
- learning process
- data sets