Autoencoding Blade Runner
Terence Broad

It would be interesting if the algorithm could be trained more on the properties of the most significant objects, like faces, in other words applying a deeper ‘plug-in’ model specifically trained to ‘know’ more about faces such as their structure from differing perspectives, during movement, in different lightening, etc. and then use that to further enhance the basic output. Pretty cool in any case, and I’ll probably dream tonight about electronically recreated sheep.

Like what you read? Give Tony Sarris a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.