Testing my understanding of Machine Learning fundamentals such as gradient descent and back propagation by creating a single-layer perceptron from scratch. Created a neural network from scratch and implemented all of the concepts from scratch, using pygame to visualize.
In this demo, the red line is the model's estimate of the gradient between the white and the black dots. The green dots are the ones that the model correctly estimated, and the red dots are incorrect. Dots that are not green nor white are currently being evaluated. I slowed the processing time for this demo so it would provide a more effective visualization of the process.