The aim of this project was to create a better model of the mouse visual system.
We did this by using a biologically inspired architecture called MouseNet that has been published here with code available here.
We used self-supervised learning algorithms for training the model. The two implementations tried are SimSiam and SimCLR. The code used for training is based on the SimSiam code available here.
For evaluating the models we used Representational Similarity Analysis (RSA) based on this repo.