Automatic tuning of hyper-parameters of reinforcement learning algorithms using Bayesian optimization with behavioral cloning.
Juan Cruz BarsceJorge A. PalombariniErnesto C. MartínezPublished in: CoRR (2021)
Keyphrases
- reinforcement learning algorithms
- hyperparameters
- posterior distribution
- bayesian inference
- reinforcement learning
- model selection
- conjugate priors
- cross validation
- maximum likelihood
- model free
- bayesian methods
- state space
- markov decision processes
- support vector
- closed form
- bayesian framework
- prior information
- parameter settings
- em algorithm
- random sampling
- learning algorithm
- sample size
- temporal difference
- incremental learning
- maximum a posteriori
- reward function
- incomplete data
- noise level
- bayesian networks
- function approximation
- missing values
- data sets
- image reconstruction
- dynamic environments
- probabilistic model
- dynamic programming
- multi agent
- multiscale
- data mining