Dynamic Sparse Training for Deep Reinforcement Learning.
Ghada SokarElena MocanuDecebal Constantin MocanuMykola PechenizkiyPeter StonePublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- learning algorithm
- training set
- high dimensional
- markov decision processes
- sparse data
- supervised learning
- dynamic environments
- deep architectures
- real time
- robot control
- temporal difference
- model free
- sparse coding
- training samples
- object detection
- bayesian networks
- image sequences
- machine learning
- neural network