Difference between revisions of "Deep Reinforcement Learning (DRL)"
(→Q=learning & SARSA) |
|||
| Line 6: | Line 6: | ||
* [http://deeplearning4j.org/deepreinforcementlearning.html Guide] | * [http://deeplearning4j.org/deepreinforcementlearning.html Guide] | ||
| − | === Q | + | === Q-learning & SARSA === |
* [[Deep Q Learning (DQN)]] | * [[Deep Q Learning (DQN)]] | ||
* [[State-Action-Reward-State-Action (SARSA)]] | * [[State-Action-Reward-State-Action (SARSA)]] | ||
* [[Neural Coreference]] | * [[Neural Coreference]] | ||
| + | |||
=== Policy Gradient Methods === | === Policy Gradient Methods === | ||
* [[Deep Deterministic Policy Gradient (DDPG)]] | * [[Deep Deterministic Policy Gradient (DDPG)]] | ||
Revision as of 21:35, 26 May 2018
- Gym | OpenAI
- Introduction to Various Reinforcement Learning Algorithms. Part I (Q-Learning, SARSA, DQN, DDPG) | Steeve Huang
- Introduction to Various Reinforcement Learning Algorithms. Part II (TRPO, PPO) | Steeve Huang
- Guide
Q-learning & SARSA
Policy Gradient Methods
- Deep Deterministic Policy Gradient (DDPG)
- Trust Region Policy Optimization (TRPO)
- Proximal Policy Optimization (PPO)
Goal-oriented algorithms, which learn how to attain a complex objective (goal) or maximize along a particular dimension over many steps; for example, maximize the points won in a game over many moves. Reinforcement learning solves the difficult problem of correlating immediate actions with the delayed returns they produce. Like humans, reinforcement learning algorithms sometimes have to wait a while to see the fruit of their decisions. They operate in a delayed return environment, where it can be difficult to understand which action leads to which outcome over many time steps.
