Understanding the Effects of Dataset Characteristics on Offline Reinforcement Learning.
Kajetan SchweighoferMarkus HofmarcherMarius-Constantin DinuPhilipp RenzAngela Bitto-NemlingVihang P. PatilSepp HochreiterPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- mechanisms underlying
- real time
- state space
- benchmark datasets
- function approximation
- relative strengths and weaknesses
- learning algorithm
- multi agent
- real life
- markov decision processes
- deeper understanding
- optimal control
- optimal policy
- action selection
- training dataset
- learning problems
- decision making
- genetic algorithm