MoËT: Mixture of Expert Trees and its application to verifiable reinforcement learning.
Marko VasicAndrija PetrovicKaiyuan WangMladen NikolicRishabh SinghSarfraz KhurshidPublished in: Neural Networks (2022)
Keyphrases
- neural network
- reinforcement learning
- multi objective
- decision trees
- function approximators
- function approximation
- reinforcement learning algorithms
- mixture model
- multi agent
- learning process
- model free
- state space
- machine learning
- markov decision processes
- tree structure
- learning algorithm
- domain experts
- expectation maximization
- differential evolution
- multi agent reinforcement learning
- gaussian distribution
- tree models
- tree structures
- temporal difference
- robotic control
- optimal control
- optimal policy
- sufficient conditions
- genetic algorithm
- expert knowledge
- gaussian mixture model
- reinforcement learning methods
- domain knowledge
- policy search