Measuring Sample Efficiency and Generalization in Reinforcement Learning Benchmarks: NeurIPS 2020 Procgen Benchmark.
Sharada P. MohantyJyotish PoonganamAdrien GaidonAndrey KolobovBlake WulfeDipam ChakrabortyGrazvydas SemetulskisJoão SchapkeJonas KubiliusJurgis PasukonisLinas KlimasMatthew J. HausknechtPatrick MacAlpineQuang Nhat TranThomas TumielXiaocheng TangXinwei ChenChristopher HesseJacob HiltonWilliam Hebgen GussSahika GencJohn SchulmanKarl CobbePublished in: NeurIPS (Competition and Demos) (2020)
Keyphrases
- reinforcement learning
- benchmark suite
- function approximation
- neural network
- real world
- highly efficient
- learning algorithm
- information retrieval
- multi agent
- computational complexity
- state space
- markov decision processes
- optimal control
- reinforcement learning algorithms
- sample points
- multi agent reinforcement learning
- database