logoalt Hacker News

WhiteNoiz3today at 1:00 PM2 repliesview on HN

With the old model (and I suspect this one too) it's trained to generate from a single 'seed' pixel in the center of the image. If you erase the center of the image, that's when it completely collapses.


Replies

oerstedtoday at 2:19 PM

It must be more general than that, otherwise the cells wouldn’t be able to repair their area if the damage came from the wrong direction (repair is not center-out).

The model generally learns to generate each pixel from its surroundings, even if the surroundings are partially missing.

show 1 reply
cl3mischtoday at 2:13 PM

Have you actually tried that? If you specifically erase the center, the image does change a lot at first, but rebuilds itself eventually (albeit to a slightly different final state). It's uncanny how "biological" is feels!

show 2 replies