Hopfield Networks and Application
This repo contains my project on "Exploring Applications of Hopfield Networks" which I presented against the certification for the course Machine Learning for Physcists.
Here I have implemented the hopfield networks from scractch in python, two specific applications explored here are that of Pattern retrieval and Combinatorial Optimization.
For Pattern retieval, I tried storing 28x28 images of handwritten digits from the MNISt dataset. Basing on this the performance of the retrieval process of the network was experimented upon for varying strength of correlation among the stored digits.
For Combinatorial optimisation, I implemented the network to solve the N-rook's problem and the validity of the produced solutions were experimented against the varying size of the game and different initial conditions.