Examples

Kohonen's SOM Neural Net Trained on Uniform Distribution

Code

Visualization via Processing (5x5 mapping neurons)

Feedforward Network Fit to Sine Function via Gradient Descent

Code

Visualization via Processing (fit to sine wave)

Proximal Policy Optimization (Rough implementation)

Code

• Policy and value functions defined separately
• Reinforcement learning framework
• Generalized Advantage Estimator implementation
• Long training time with current state (no multiprocessing)