Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Feature Visualization: How neural nets build up their understanding of images

distill.pub

Post reply on HN
Print view View original

51–60 of 66 posts

Re: Feature Visualization: How neural nets build up their understanding of images

#51

Post by cm2187 »

Houshalter wrote: post #49
2bitencryption wrote: post #46

Earlier quoted context omitted.

the school bus one really sticks out to me. it seems all the net cares about is seeing orange juxtaposed with black. No shapes, no vehicle features, just orange and black. hazarding a guess with no knowledge of the subject, I wonder if that is because no other class in ImageNet can be defined by orange and black. The net simply doesn't need to learn anything more about orange and black, because on 100% of the samples…

That's a good observation. But when I see that image I definitely think of a school bus. The color pattern is very distinctive (in the US anyway where all school buses are painted the same color and style.) So I can't say the NN is wrong. It doesn't necessarily mean that color is the only feature it uses to classify school buses. Just that that feature alone is enough.

would be interesting to try these pictures, the stripes would compete with the other features of the picture, so it might not result in seeing a schoolbus.

https://static.comicvine.com/uploads/scale_large/6/67663/418...

http://www.fairmont.com/assets/0/137/13359/13476/13477/2a657...

Re: Feature Visualization: How neural nets build up their understanding of images

#53

Post by Scarblac »

colah3 wrote: post #2

Hey! I'm one of the authors, along with Alex and Ludwig. We're happy to answer any questions! :)

Do you often have beautiful pictures like this on your screen?

(a reason why I'm moving into scientific visualization and GIS instead of backend work...)

Re: Feature Visualization: How neural nets build up their understanding of images

#55

Post by taneq »

chillingeffect wrote: post #19

Is there any way to run images from a camera real-time into GoogLeNet? E.g. like if I want to scan areas around me to see if there are any perspectives in my environment that light up the "snake" neurons or the dog neurons???

If not in realtime, you could always take a video and post-process it offline?

Re: Feature Visualization: How neural nets build up their understanding of images

#56

Post by zkms »

shellbackground wrote: post #40

This pictures reminds me about what one's can see under psychedelics. All sensory input basically begins to break down to that kind of patterns, and thus reality dissolves into nothing. This is equally terrifying and liberating depends on look. The terrifying thought is that there's no-one behind this eyes and ears. The liberating thought is that if there's no-one there, then there's no-one to die.

> This pictures reminds me about what one's can see under psychedelics. All sensory input basically begins to break down to that kind of patterns

The neural structures in human brains that recognise edges/textures/surfaces are the same ones that generate trippy images when exposed to psychedelic drugs (or flickering light: http://journals.plos.org/ploscompbiol/article?id=10.1371/jou...): http://www.math.utah.edu/~bresslof/publications/01-3.pdf

Re: Feature Visualization: How neural nets build up their understanding of images

#57

Post by matt4077 »

shellbackground wrote: post #40

This pictures reminds me about what one's can see under psychedelics. All sensory input basically begins to break down to that kind of patterns, and thus reality dissolves into nothing. This is equally terrifying and liberating depends on look. The terrifying thought is that there's no-one behind this eyes and ears. The liberating thought is that if there's no-one there, then there's no-one to die.

have you seen "Deep Dream"? https://www.youtube.com/watch?v=SCE-QeDfXtA and https://www.youtube.com/watch?v=DgPaCWJL7XI are excellent examples.

I believe for anybody who has ever tried psychedelics seeing these was a watershed moment. It seems almost impossible that an algorithm could so faithfully reproduce the experience without somehow having recreated some fundamental structure of the human brain. That is compounded by the fact that this wasn't the result of actually trying to do so, but of an open-ended experiment of creating feedback loops in NN layers.

Re: Feature Visualization: How neural nets build up their understanding of images

#59

Post by natch »

shancarter wrote: post #3

There’s also an appendix where you can browse all the layers. https://distill.pub/2017/feature-visualization/appendix/goog...

Yes that appendix is not to be missed. Also clicking on an image goes deeper into an exhaustive catalog of each layer, which is an amazing and huge resource.

Re: Feature Visualization: How neural nets build up their understanding of images

#60

Post by colah3 »

Scarblac wrote: post #53
colah3 wrote: post #2

Hey! I'm one of the authors, along with Alex and Ludwig. We're happy to answer any questions! :)

Do you often have beautiful pictures like this on your screen? (a reason why I'm moving into scientific visualization and GIS instead of backend work...)

Yep, neural network interpretability research involves quite a bit of staring a pretty pictures. It's a difficult job, but someone has to look at all those images. :P
Post reply on HN