Generative Adversarial Networks – The Story So Far • Hacker News Board

Re: Generative Adversarial Networks – The Story So Far

#11

Post by gwern » Sun, Jun 23, 2019, 1:42 AM UTC

heydenberk wrote: post #8

6gvONxR4sf7o wrote: post #7

Earlier quoted context omitted.

I know people have been having trouble adapting these kinds of generative techniques to text. Do you know of anyone making interesting progress there?

I put your question in talktotransformer.com and got this response: I know people have been having trouble adapting these kinds of generative techniques to text. Do you know of anyone making interesting progress there? It is difficult to make progress in the field of generative methods with text alone - it takes effort and creativity to get a generative system working. A big part of our research focuses on generating…

GPT-2 is, however, not adversarial at all, and that might be part of why it rambles and lacks consistency or much of a 'point'.

Re: Generative Adversarial Networks – The Story So Far

#12

Post by mkl » Sun, Jun 23, 2019, 1:56 AM UTC

galaxyLogic wrote: post #9

This basically means you can take any politician or citizen and create a video in which they say or do bad (but fake) things. Age of Fake News has truly arrived. Is there some research on how to distinguish fake generated videos from real ones?

The whole idea of Generative Adversarial Networks is that you're simultaneously training a network that can generate fake things, and another network that can distinguish fake from real. The two networks are each other's adversary, the generator trying to fool the discriminator, and the discriminator trying to catch the generator.

The article explains this very accessibly.

Re: Generative Adversarial Networks – The Story So Far

#13

Post by Jedi72 » Sun, Jun 23, 2019, 2:37 AM UTC

mkl wrote: post #12

galaxyLogic wrote: post #9

This basically means you can take any politician or citizen and create a video in which they say or do bad (but fake) things. Age of Fake News has truly arrived. Is there some research on how to distinguish fake generated videos from real ones?

The whole idea of Generative Adversarial Networks is that you're simultaneously training a network that can generate fake things, and another network that can distinguish fake from real. The two networks are each other's adversary, the generator trying to fool the discriminator, and the discriminator trying to catch the generator. The article explains this very accessibly.

It also explains that the training is basically done when the discriminator has a roughly 50% (random) ability to discern between real and fake.

Re: Generative Adversarial Networks – The Story So Far

#14

Post by gwern » Sun, Jun 23, 2019, 2:38 AM UTC

hooloovoo_zoo wrote: post #4

I wonder how these GANs can be improved to have better symmetry. All these faces seem to have mismatched eyes/ears which must be coming from the locality of CNNs.

Self-attention seems to help a lot with that. (My usual example is that with anime faces, GANs which don't use either self-attention or progressive growing frequently have a failure mode of mismatched eye color: red+blue, for example.) Self-attention is expensive, so even BigGAN uses them lightly, typically only once, like at the 64px level, but there's work on making self-attention a lot cheaper and approaches like the Sparse Transformer (which OA recently used for MuseNet to let the Transformer scale to 30k-long sequences) are promising for making self-attention a lot cheaper.

Re: Generative Adversarial Networks – The Story So Far

#15

Post by mkl » Sun, Jun 23, 2019, 2:45 AM UTC

Jedi72 wrote: post #13

mkl wrote: post #12

Earlier quoted context omitted.

The whole idea of Generative Adversarial Networks is that you're simultaneously training a network that can generate fake things, and another network that can distinguish fake from real. The two networks are each other's adversary, the generator trying to fool the discriminator, and the discriminator trying to catch the generator. The article explains this very accessibly.

It also explains that the training is basically done when the discriminator has a roughly 50% (random) ability to discern between real and fake.

True. I'm not sure "done" is quite the right word. I think the problem is at that point you no longer have a useful training signal, so you have to stop.

Re: Generative Adversarial Networks – The Story So Far

#16

Post by tachyonbeam » Sun, Jun 23, 2019, 3:30 AM UTC

amelius wrote: post #10

iyaja wrote: post #2

Hi everyone. I just published a new blog post which talks about the evolution of GANs over the last few years. You can check it out here. I think it's fascinating to see sample images generated from these models side by side. It really does give a sense of how fast this field has progressed. In just five years, we've gone from blurry, grayscale pixel arrays that vaguely resemble human faces to thispersondoesnotexist,…

Cool! But one thing I'd like to see discussed is to what extent the images in various publications have been cherry-picked.

I'm curious about that too, and I'd like to know if there's been much work on GANs that can generate videos.

Re: Generative Adversarial Networks – The Story So Far

#17

Post by lettergram » Sun, Jun 23, 2019, 4:25 AM UTC

6gvONxR4sf7o wrote: post #7

iyaja wrote: post #2

Hi everyone. I just published a new blog post which talks about the evolution of GANs over the last few years. You can check it out here. I think it's fascinating to see sample images generated from these models side by side. It really does give a sense of how fast this field has progressed. In just five years, we've gone from blurry, grayscale pixel arrays that vaguely resemble human faces to thispersondoesnotexist,…

I know people have been having trouble adapting these kinds of generative techniques to text. Do you know of anyone making interesting progress there?

Yes!

https://medium.com/capital-one-tech/why-you-dont-necessarily...

GANs simply try to replicate a set of features - you can think of this as images or text. Variations in the GAN designs will be present, but the general principles are the same.

Re: Generative Adversarial Networks – The Story So Far

#18

Post by TomMarius » Sun, Jun 23, 2019, 6:59 AM UTC

Jedi72 wrote: post #13

mkl wrote: post #12

Earlier quoted context omitted.

The whole idea of Generative Adversarial Networks is that you're simultaneously training a network that can generate fake things, and another network that can distinguish fake from real. The two networks are each other's adversary, the generator trying to fool the discriminator, and the discriminator trying to catch the generator. The article explains this very accessibly.

It also explains that the training is basically done when the discriminator has a roughly 50% (random) ability to discern between real and fake.

You can simply use the detector network to discard faulty results in a pipeline and it's already useful

Re: Generative Adversarial Networks – The Story So Far

#19

Post by Scaevolus » Sun, Jun 23, 2019, 7:01 AM UTC

amelius wrote: post #10

iyaja wrote: post #2

Hi everyone. I just published a new blog post which talks about the evolution of GANs over the last few years. You can check it out here. I think it's fascinating to see sample images generated from these models side by side. It really does give a sense of how fast this field has progressed. In just five years, we've gone from blurry, grayscale pixel arrays that vaguely resemble human faces to thispersondoesnotexist,…

Cool! But one thing I'd like to see discussed is to what extent the images in various publications have been cherry-picked.

Cherry picking is no longer necessary with recent advancements. The images on https://thispersondoesnotexist.com/ are random and have a few artifacts (particularly the backgrounds and hair), but if you weren't looking for it you'd be unlikely to notice anything.

Re: Generative Adversarial Networks – The Story So Far

#20

Post by Erlich_Bachman » Sun, Jun 23, 2019, 7:34 AM UTC

galaxyLogic wrote: post #9

This basically means you can take any politician or citizen and create a video in which they say or do bad (but fake) things. Age of Fake News has truly arrived. Is there some research on how to distinguish fake generated videos from real ones?

You could produce very good photoshop fakes like these for a very long time now. It has had basically no effect on the politics and everyone knows that this technology exists and treats all the pictures they see accordingly (checking their sources etc). It will just be the same with video.