mindmap
root((RL))
Main Concepts
[MDP]
Continuing Tasks
Episodic Tasks
Markov Property
[Reward]
[Action Values]
[Policy]
{{maps states to likelihood of actions}}
Deterministic
{{one action per state}}
Stochastic
{{multiple actions per state}}
Exploration
Exploitation
[Policy Ealuation - Predication]
{}
Control
{}
Dynamic programming
Synchronous
Asynchronous
Learning
On Policy learning
{{Agents learn from their policy}}
Off Policy learning
{{Agents lean from another policy or Data}}
Online
Offline
Optimistic initial values
Math
Bellman Equations
{{State-Value Function}}
{{Action-Value Function}}
{{State-Value Optimality Function}}
{{Action-Value Optimality Function}}
Policy Improvement Theorem
Algorithms
[Bandits]
Epsilon greedy
Thompson sampling
Upper confidence bound
Contextual
Regret
Follow the normalized leader
Contractual regret
Greedyfication
[Policy Iteration]
[Generalized Policy Iteration]
[Value Iteration]
Brute force search
Monte Carlo
Bootstrapping
[Sample Based Methods]
[Temporal Difference Learning]
[SARSA]
[Q-Learning]
[Function Approximation Methods]
Others
Dyna
Citation
BibTeX citation:
@online{bochman2024,
author = {Bochman, Oren},
title = {RL {MindMap}},
date = {2024-12-20},
url = {https://orenbochman.github.io/posts/2024/2024-03-25-rl-maps/},
langid = {en}
}
For attribution, please cite this work as:
Bochman, Oren. 2024. “RL MindMap.” December 20, 2024. https://orenbochman.github.io/posts/2024/2024-03-25-rl-maps/.