Q-learning-like RL algorithm that utilizes vector-symbolic architecture and hypervectors. Solves the cart-pole environment. Now includes Double-Q learning and a target model to improve the stability of the reward convergence. This code routinely converges to > 200 iterations upright well within 100 about episodes. Max performance goes into the tens of thousands of iterations upright.
-
Notifications
You must be signed in to change notification settings - Fork 0
HowardGoldowsky/Vector_symbolic_RL
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Q-learning-like RL algorithm that utilizes vector-symbolic architecture and hypervectors.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published