Sometimes on Reddit you can accidentally venture into the far reaches of the internet and discover a whole new way of looking at the world. Such was my journey this evening when I found Reddit’s Deep Dream Images page. Not a somnambulist’s journal, but rather a foray into machine learning where intelligent computers on a rapid trajectory towards sentience, add new complex layers of understanding to photographic images. Still confused? Here’s an internet definition of deep dream images and deep learning.
Deep Learning is a new field within Machine Learning. In the past 4 years researchers have been training neural networks with a very large number of layers. Algorithms are learning how to classify images to a much greater accuracy than before: you can give them an image of a cat or a dog and they will be able to tell the difference. Traditionally this has been nearly impossible for computers but easy for humans.
Deep learning algorithms are trained by giving them a huge number of images, and telling them what object is in each image. Once it has seen (e.g.) a hundred types of dog heads 1000 times from a hundred angles, it has been ‘trained’. Now you can give it new images and it will spot dog heads within the images, or tell you that there are none at all. It also can say how unsure it is.
It was always hard to tell what the algorithms were ‘seeing’ or ‘thinking’ when we gave them new images. So in June 2015 Google Engineers released a method for visualising what the algorithms saw.. Towards the end of June 2015 they released their code, so people could see what the trained neural networks were seeing on any image they wanted.
Best of all, you can make your own kooky, mind-bending images using this software online:
Try it yourself with the deep dream generator
Let me know how you go…
Here’s a little something to blow your mind as well. Who needs to smoke anything nowadays!