Challenges of Real-World Reinforcement Learning.
Gabriel Dulac-ArnoldDaniel J. MankowitzTodd HesterPublished in: CoRR (2019)
Keyphrases
- real world
- reinforcement learning
- wide range
- lessons learned
- synthetic data
- technical challenges
- dynamic programming
- data sets
- case study
- real life
- function approximation
- open issues
- database
- robotic control
- real robot
- function approximators
- temporal difference
- least squares
- key issues
- learning problems
- markov decision processes
- state space
- multiscale
- social networks
- databases
- real time