RL Theory - Githubissues

eleurent / phd-bibliography

References on Optimal Control, Reinforcement Learning and Motion Planning

902 stars 205 forks source link

RL Theory #1

Open eleurent opened 4 years ago

eleurent commented 4 years ago

RL Theory is not properly represented. A new section should be added, with at least:

Tabular setting
- With a generative model
- QVI
- Without
- UCRL2
- UCBVI
- Episodic
- Q-learning+UCB
Extensions to compact state-action spaces
Extension to Kernels
Performance measures: PAC, simple regret, cumulative regret, etc.
RL with compatible function approximation

Is there a difference between generative models (sample any transition) and simulators (simulate trajectories from current states only)?