So I whipped up a couple webpages to track the progress of how the evolved convolutional neural networks are doing and what they're looking like and thought you might like to see a little bit beneath the hood about what's going on as well.
This page displays an overview of each search that's being run:
You can click on the search names to get a more detailed view of how that search is performing, e.g.:
And then click on a genome to see what that convolutional neural network actually looks like:
Some of them are pretty interesting! The best networks we've evolved so far are already above 95% accuracy on the training data from the MNIST handwritten digits data set which is getting pretty good!
If you have any questions about what you're looking at and want some more detailed explanation please ask and I'll do what I can, as well as putting more information on the webpages to make them more volunteer friendly.