So in concrete terms, does this mean that we show the network an image, choose o... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

gradys on June 18, 2015 | parent | context | favorite | on: Inceptionism: Going Deeper into Neural Networks

So in concrete terms, does this mean that we show the network an image, choose one layer's output vector, and then back-propagate gradients to the image such that the direction of that vector stays the same, but the magnitude increases?

murbard2 on June 18, 2015 [–]

That is my understanding of the blog post, yes.

That plus a prior on the input pixels to keep it image-like.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact