Predictive Coding for Boosting Deep Reinforcement Learning with Sparse Rewards.
Xingyu LuStas TiomkinPieter AbbeelPublished in: CoRR (2019)
Keyphrases
- reinforcement learning
- predictive coding
- lossless compression
- learning algorithm
- binary tree
- motion compensated
- entropy coding
- markov decision processes
- concept drift
- video coding
- machine learning
- non stationary
- reward function
- high dimensional
- data sets
- state space
- optimal policy
- image compression
- compression scheme
- transform coding
- compression ratio
- image processing
- feature selection
- sparse representation
- supervised learning
- arithmetic coding