I've updated the algorithm behind neuralgae. It's still the same basic principle: render a deepdraw image based on some random categories, feed the results back through the classifier, and then repeat with the new categories. But the draw part of the cycle now uses the weights from the classification stage. I also made the background images a bit more varied and colourful.
There's a little less randomness, and when it gets stuck in a big set of subcategories like dog breeds it can still get completely bogged down, but when it's good it produces exactly the sort of dreamlike wandering from one surreal image to the next which I was aiming at.
The new-style images are appearing on the Twitter feed now.