The Importance of Pessimism in Fixed-Dataset Policy Optimization.
Jacob BuckmanCarles GeladaMarc G. BellemarePublished in: CoRR (2020)
Keyphrases
- optimization algorithm
- optimization process
- relative importance
- optimal policy
- discrete optimization
- optimization method
- machine learning
- global optimization
- benchmark datasets
- optimization problems
- evolutionary algorithm
- database
- combinatorial optimization
- information technology
- training dataset
- reinforcement learning
- asymptotically optimal
- direct search