Computer generated sexy stories, all you do is give it an image to work with…

Samim Winiger, a Code magician and Narrative engineer, has used an open-source neural network that was trained on 14 million passages of romance novels by Ryan Kiros, a University of Toronto PhD student specializing in machine learning.

Aptly titled the Neural-Storyteller, the network was trained to analyze images and retrieve appropriate captions from its vast store of sexy knowledge, creating “little stories about images,” says Kiros.

Here’s a sample.

neural net 2

More on this story here

Open-source code for the Neural-Storyteller here

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.