03 30th January, Tuesday

Written a semi-functional implementation of t-SNE for analysing the weights in the last layer of the network of my traffic sign recognition network from the ADL project last term. Seems as though tomorrow I should be able to get it fully functional. I'm hoping that this will give some important insight into detecting the attacks.

I have now run the histogram analyser against the sign encoding attack, but have now realised that my implementation of Resnet is too small to convey the information that I want to receive, so training a larger network for more epochs will probably be necessary.

I need to write a short abstract of the project in order to apply for a project using Blue Crystal. I am planning on emailing this to Professor Flach tomorrow so that he can apply on my behalf (I've heard that requesting your supervisor apply for you is the way to apply, but cannot confirm).