MoVEMo: A Structured Approach for Engineering Reward Functions.
Piergiuseppe MallozziRaúl PardoVincent DuplessisPatrizio PelliccioneGerardo SchneiderPublished in: IRC (2018)
Keyphrases
- reward function
- markov decision processes
- reinforcement learning
- state space
- multiple agents
- inverse reinforcement learning
- transition probabilities
- simple examples
- optimal policy
- state variables
- policy search
- markov decision process
- markov models
- machine learning
- structured data
- reinforcement learning algorithms
- random walk
- dynamic programming
- active learning
- learning algorithm