Neural Networks with Few Labels
This was a class project for the coursework of Computer Vision. It was presented in the form of an assignment split into two parts- 1) Derive the backpropagation algorithm given the loss function of the network 2) Implement the backpropagation in MATLAB.
Background of problem can be found in the uploaded pdf. The backpropagation is derived step by step in the corresponding pdf. Most of the code was given by the instructor. My task was to implement the gradient computation and weight updation.