Evolutionary reinforcement learning for sparse rewards.
Shibei ZhuFrancesco BelardinelliBorja Gonzalez LeónPublished in: GECCO Companion (2021)
Keyphrases
- reinforcement learning
- function approximation
- markov decision processes
- genetic algorithm
- state space
- learning algorithm
- model free
- reinforcement learning algorithms
- machine learning
- sparse representation
- evolutionary computation
- multi agent
- evolutionary optimization
- sparse data
- reward function
- temporal difference
- transfer learning
- optimal policy
- high dimensional
- policy iteration
- robotic control
- reward shaping
- total reward
- hidden state
- sparse coding
- dynamic programming
- markov decision process
- compressed sensing
- complex domains
- action space
- feature space
- reinforcement learning methods
- sparse matrix
- markov decision problems
- partially observable
- multi agent reinforcement learning
- compressive sensing
- data sets
- multiarmed bandit