Google Brain's Magenta: Multi-Style Image Transfer with Code

As an artist I find it very frustrating when people try to apply style transfer type techniques in an attempt to emulate an artist like Picasso. It kinda works and generates a bunch of hype but it isn&#x27;t even close. The reason it&#x27;s frustrating is because I think Deep Learning is <i>actually</i> capable of doing this stuff but the people implementing need to understand how Picasso actually did his work.<p>If you look at cubism the whole idea is to capture multiple sides of a 3-Dimensional object at once. A lot of art is not a &quot;style&quot; but rather a projection from 3D (or 4D) space to 2D space.<p>If you wanted to paint a &quot;dog&quot; in the style of Picasso your network would need to understand the geometry of a dog.<p>Training on a bunch of 2D before and after training examples is underspecified.<p>It&#x27;s important to understand that it is a mapping from 3D -&gt; 2D ... <i>NOT</i> 2D-&gt;2D<p>Another example is &quot;Nude Descending a Staircase&quot; by Duchamp: <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Nude_Descending_a_Staircase,_No._2" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Nude_Descending_a_Staircase,_N...</a><p>It is a painting describing motion. To apply style transfer would be completely stupid because the point of the image is to project 4D-&gt;2D ... not to have wavy black and brown lines.

Why is everyone working on style transfer? It doesn&#x27;t seem like such an interesting problem in the field, compared to things like speech recognition for example. Is it just because it&#x27;s a &quot;cracked&quot; problem and it looks nice? I&#x27;m just genuinely curious here, not trying to bash the amazing work these people do.

Not sure how related this is but seems like the right crowd to ask...<p>As a photographer who also does programming full time I&#x27;ve been wondering what would need to happen to synthesize skin texture to remove imperfections. Example, removing small scars, wrinkles, etc. Currently I just use the healing brush in Photoshop but wondering if ML can be used to automatically do it.<p>Does anyone have any recommendations on what sub-fields or papers I could read to get a better idea of what would be involved to create a solution like that?