Regularization and Variance-Weighted Regression Achieves Minimax Optimality in Linear MDPs: Theory and Practice.
Toshinori KitamuraTadashi KozunoYunhao TangNino VieillardMichal ValkoWenhao YangJincheng MeiPierre MénardMohammad Gheshlaghi AzarRémi MunosOlivier PietquinMatthieu GeistCsaba SzepesváriWataru KumagaiYutaka MatsuoPublished in: ICML (2023)
Keyphrases
- square loss
- aggregating algorithm
- simple linear
- kernel ridge regression
- model selection
- markov decision processes
- regression model
- ridge regression
- search algorithm
- reproducing kernel hilbert space
- reinforcement learning
- linear models
- nonlinear regression
- average cost
- linear regression
- kernel methods
- state space
- gaussian processes
- regularization parameter
- variance reduction
- regularization framework
- sufficient conditions
- gradient boosting
- factored mdps
- support vector
- optimal solution