Exploration-Guided Reward Shaping for Reinforcement Learning under Sparse Rewards.
Rati DevidzeParameswaran KamalarubanAdish SinglaPublished in: NeurIPS (2022)
Keyphrases
- reward shaping
- reinforcement learning
- reinforcement learning algorithms
- complex domains
- action selection
- state space
- function approximation
- markov decision problems
- multi agent
- model free
- dynamic programming
- machine learning
- optimal control
- learning algorithm
- neural network
- temporal difference
- objective function
- learning agent
- function approximators
- continuous state