The Curse of Passive Data Collection in Batch Reinforcement Learning.
Chenjun XiaoIlbin LeeBo DaiDale SchuurmansCsaba SzepesváriPublished in: AISTATS (2022)
Keyphrases
- data collection
- reinforcement learning
- high dimensional
- batch mode
- data analysis
- collected data
- function approximation
- sensor networks
- markov decision processes
- state space
- high dimensional data
- collecting data
- temporal difference
- dimension reduction
- dimensionality reduction
- multi agent
- reinforcement learning algorithms
- robotic control
- control policy
- learning algorithm
- optimal policy
- database
- action selection
- optimal control
- action space
- reinforcement learning methods
- data entry
- wireless sensor networks
- multi agent reinforcement learning
- policy search
- high dimensionality
- online learning
- temporal difference learning
- quality prediction
- batch size
- batch learning