When Should We Prefer Offline Reinforcement Learning Over Behavioral Cloning?
Aviral KumarJoey HongAnikait SinghSergey LevinePublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- model free
- function approximation
- reinforcement learning algorithms
- real time
- multi agent
- selective perception
- robot control
- markov decision processes
- state space
- machine learning
- policy search
- learning algorithm
- learning problems
- genetic algorithm
- robotic control
- perceptual aliasing
- neural network
- partially observable
- databases
- multi agent reinforcement learning
- evolutionary learning
- agent behavior
- control problems
- artificial intelligence
- real world
- reward function
- human behavior
- artificial neural networks