NeoRL: A Near Real-World Benchmark for Offline Reinforcement Learning.
Rongjun QinSongyi GaoXingyuan ZhangZhen XuShengkai HuangZewen LiWeinan ZhangYang YuPublished in: CoRR (2021)
Keyphrases
- real world
- reinforcement learning
- function approximation
- synthetic data
- case study
- wide range
- multi agent
- reinforcement learning algorithms
- comparative analysis
- markov decision processes
- data structure
- real life
- state space
- feature selection
- search engine
- learning algorithm
- data mining
- real time
- database systems
- multi agent systems
- monte carlo
- model free