The Graphic City: The winner of the IBTimes UK minimalist architecture photography competition

The winners of the photography competition we ran in conjunction with EyeEm, using machine learning and neural networks to pick the winners.

In February, International Business Times UK teamed up with EyeEm for a photography competition with a difference: one judged by Artificial Intelligence. EyeEm's patented technology uses machine learning and neural networks to identify not only what of a photograph is of, but also whether it is any good.

The theme we chose was 'The Graphic City'. We wanted photographers to visualise the lines, curves and corners of cities. We were looking for interesting angles, converging lines, and shapes revealed in creative compositions of minimalist architecture across the globe.

We were quite excited to see not only what EyeEm's photographers would come up with, but also whether we would agree with the computer's choices. It turns out we did. We received more than 46,000 entries, and can now reveal the results.

Our overall winner, as chosen by the machines (and agreed upon by a few humans) is American photographer Jeremy Walter, who had several entries in the top 50 images. Jeremy says he is "insatiably attracted to architecture, particularly the geometry of patterns and forms inherent in that field. Simple lines and curves offer worlds of wonders to me. I love discovering things that already exist: hidden symphonies of shapes that mute the world around me."

These are some of Jeremy's entries which caught the computer's (virtual) eye. IBTimes UK has now published a standalone gallery of his favourite images.

EyeEm regularly invites its global community of more than 20 million photographers to partake in competitions, or 'Missions'. These attract tens of thousands of entries, so it would be a Herculean task for a human judge to even look at every image, let alone make value judgements. However, their AI does this in seconds.

EyeEm's technology can also be trained to find photos with a particular aesthetic. Feed it a small batch of images, and it will search its vast database for other images with similar attributes. The remarkable thing is that the images it chooses don't all end up looking identical. Here are some of the other shortlisted images from our competition 'The Graphic City'. They all fit the brief, but they're all quite different.