An empirical investigation of the challenges of real-world reinforcement learning.
Gabriel Dulac-ArnoldNir LevineDaniel J. MankowitzJerry LiCosmin PaduraruSven GowalTodd HesterPublished in: CoRR (2020)
Keyphrases
- real world
- reinforcement learning
- wide range
- key issues
- lessons learned
- data sets
- temporal difference
- model free
- learning algorithm
- state space
- application scenarios
- action selection
- function approximation
- case study
- genetic algorithm
- data mining
- neural network
- real time
- hidden markov models
- information systems
- optimal control
- machine learning
- technical challenges
- learning capabilities
- open issues
- temporal difference learning