Building neural nets from scratch, utilising back propagation and derivative sigmoid model for 2 Layer neural nets while using deep learning to develop a 3 layer model for solving more complex binary coded puzzles.
A bare bones neural network implementation to find out the inner workings of backpropagation.