Data-efficient Deep Reinforcement Learning Method Toward Scaling Continuous Robotic Task with Sparse Rewards.
Junkai RenYichuan ZhangYujun ZengYixing LanPublished in: RCAR (2021)
Keyphrases
- synthetic data
- reinforcement learning
- test data
- prior knowledge
- database
- statistical methods
- input data
- objective function
- policy search
- pairwise
- data sets
- data collection
- computationally efficient
- training samples
- prior information
- training data
- data sources
- sparse data
- information loss
- face recognition
- probability distribution
- bayesian networks
- clustering method
- function approximation
- data analysis
- model free
- data points