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

Do vision transformers see like convolutional neural networks?

arxiv.org

Post reply on HN
Print view View original

41–45 of 45 posts

Re: Do vision transformers see like convolutional neural networks?

#41

Post by space_fountain »

spullara wrote: post #32
stared wrote: post #13

I am much more interested if they fall for the same tricks. For example, if it is easy to fool them with optical illusions, such as innocent images that look racy at the first glance: https://medium.com/@marekkcichy/does-ai-have-a-dirty-mind-to... CW: Even though it does not contain a single explicit picture, it might be considered NSFW (literally - as at the first glance it looks like nudity); full disclosure: I men…

Since it easy to fool people with optical illusions I doubt that you will be able to train a computer to not be fooled by optical illusions.

I think because the human world is made for humans there’s a lot of value in an AI with similar failure modes to humans. Right now AI can do a good job of learning to classify images, but fails in ways that are entirely foreign to us

Re: Do vision transformers see like convolutional neural networks?

#42

Post by mannykannot »

space_fountain wrote: post #41
spullara wrote: post #32

Earlier quoted context omitted.

Since it easy to fool people with optical illusions I doubt that you will be able to train a computer to not be fooled by optical illusions.

I think because the human world is made for humans there’s a lot of value in an AI with similar failure modes to humans. Right now AI can do a good job of learning to classify images, but fails in ways that are entirely foreign to us

We can use our understanding of how the world usually is and plausibly might be to get us out of local minima.

Re: Do vision transformers see like convolutional neural networks?

#43

Post by IIAOPSW »

spullara wrote: post #32
stared wrote: post #13

I am much more interested if they fall for the same tricks. For example, if it is easy to fool them with optical illusions, such as innocent images that look racy at the first glance: https://medium.com/@marekkcichy/does-ai-have-a-dirty-mind-to... CW: Even though it does not contain a single explicit picture, it might be considered NSFW (literally - as at the first glance it looks like nudity); full disclosure: I men…

Since it easy to fool people with optical illusions I doubt that you will be able to train a computer to not be fooled by optical illusions.

The hard thing to emulate is that people quickly become aware that they are looking at an illusion. Even though you can't turn your perception off, Escher's infinite staircase doesn't actually trick you into thinking a set of stairs can go in a closed loop.

Re: Do vision transformers see like convolutional neural networks?

#44

Post by Dylan16807 »

codeflo wrote: post #35
codeflo wrote: post #27

Earlier quoted context omitted.

Those pictures are definitely NSFW when viewed at low res/from far away, which is how coworkers typically see your monitor contents. An argument that starts with “Well, technically” is unlikely to carry much weight in a discussion with HR (and probably rightfully so).

Counter argument in addition to downvote, please?

Probably because "Well, technically" is a really unfair way to characterize that argument.

It's not porn. It's not simulated porn. It's a hallway, and if you're not setting it as your desktop background to trick people on purpose then you're not doing anything wrong.

Re: Do vision transformers see like convolutional neural networks?

#45

Post by sdenton4 »

kettleballroll wrote: post #2

It's always nice to see big labs working more towards building an understanding of things instead of just chasing SOTA. On the other hand, I'm not sure there is a lot of actionable findings in here. I guess that's the trade off with these things though....

Demonstrating really different representations is a good indicator that ensembles are worth trying.
Post reply on HN