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

Stable Attribution

stableattribution.com

Post reply on HN
Print view View original

361–365 of 365 posts

Re: Stable Attribution

#361

Post by whywhywhywhy »

pksebben wrote: post #340
whywhywhywhy wrote: post #302

Earlier quoted context omitted.

It’s insane how much of a gift to artists Stable Diffusion is, this tech could have been wrapped up in extremely expensive subscriptions by any of the current creative tool rent barons but it was handed to us all for free. Many are too angry and small minded to see how lucky we are. A tool beyond anything Adobe currently ships today, for free. Could have easily been a AutoCad level subscription (10k+ a year IIRC that…

I think that some of the friction here is about mindset. I have a lot of friends who are artists, and many of them consider this stuff 'tech shit I'll never understand'; I think, because we've culturally set ourselves up in a strata of 'professionals' (who are assumed to be better / smarter / uniquely capable) and non-professionals (who are born without whatever it takes to grok x ). This is along many lines - my art…

Future of creativity and artistry isn't just being expert in one thing anyway, already seeing in a lot of fields where top level creatives are managing to excel in several fields.

These sort of tools are only going to accelerate this trend in my eyes.

Re: Stable Attribution

#362

Post by intrasight »

renonce wrote: post #222
intrasight wrote: post #215

As one of my best friends told me in 1971 (we were six!), every image and sound that we produce has already been produced somewhere else in the infinite universe.

The observable universe has only 10^80 atoms. A small image of 128x128 pixels has more variations than that.

The two measure have nothing to do with each other.

Re: Stable Attribution

#363

Post by temp512345 »

pksebben wrote: post #360
dahart wrote: post #357

Earlier quoted context omitted.

This is dismissive in the face of increasing evidence that a bunch of NN models have already been caught reproducing accidentally overfit data. Many examples have popped up with Stable Diffusion, not just one you disagree with. Same goes for ChatGPT, for GitHub Copilot, for Imagen, and a bunch of models. Calling people dumb is to be willfully ignorant to the fact that neural networks actually can and really do rememb…

I agree with you that they have the capacity to remember an image - but they're not compressing them. That's a fundamentally different thing. The argument being made by that class action lawsuit is that "this thing can reproduce image X so it's a compression algorithm and nothing more", which they are predicating on an exercise that is sneaky and dishonest, and only likely to hold water with someone who has a limited…

Compression (especially a lossy one) means storing a smaller sample of the original data in whatever form you desire and then using some algorithm to reconstruct the original data up to some acceptable approximation. I would argue that in the situation we are discussing the network does just that and it is obvious to everyone involved.

Re: Stable Attribution

#364

Post by pksebben »

temp512345 wrote: post #363
pksebben wrote: post #360

Earlier quoted context omitted.

I agree with you that they have the capacity to remember an image - but they're not compressing them. That's a fundamentally different thing. The argument being made by that class action lawsuit is that "this thing can reproduce image X so it's a compression algorithm and nothing more", which they are predicating on an exercise that is sneaky and dishonest, and only likely to hold water with someone who has a limited…

Compression (especially a lossy one) means storing a smaller sample of the original data in whatever form you desire and then using some algorithm to reconstruct the original data up to some acceptable approximation. I would argue that in the situation we are discussing the network does just that and it is obvious to everyone involved.

this is the crux of my issue with the term "compression" in this context: Is it a smaller version of the data?

Yes, the model is smaller than the total input data. But when it comes to recreating a single image, how many of the weights must be configured 'just-so' recreate an image enough to call it the same image? I'll admit ignorance here - but I also don't think that this is a thing anyone knows for sure. We can only just extract othello piece colors from a simplified, othello-specialized model designed to recognize two colors.

How much of the information from other images must be present to perform this task?

My instinct, given my understanding of how these things work, is that to replicate an image with any recognizable fidelity, you have to overtrain the model enough that you've affected a set of weights much, much larger than the pixel data. The internal representation of these images is concerned with much more visual information than just 'this pixel is this color' - by looking at layered outputs from the inverse type of system (image recognition, which is the core component of these models), you can see that they're encoding layers of shading, lines that map to brushstrokes or object boundaries, foreground, background, all kinds of stuff. A direct representation of an image with all of these would be necessarily huge - and we know this because we have them. Artists use layers in all kinds of image-creation software, and they're always way bigger than the JPEG itself.

I get that this may sound pedantic, but the term 'compression' doesn't seem, to me, that it fits here. Compression, by definition, makes stuff smaller

Re: Stable Attribution

#365

Post by temp512345 »

pksebben wrote: post #364
temp512345 wrote: post #363

Earlier quoted context omitted.

Compression (especially a lossy one) means storing a smaller sample of the original data in whatever form you desire and then using some algorithm to reconstruct the original data up to some acceptable approximation. I would argue that in the situation we are discussing the network does just that and it is obvious to everyone involved.

this is the crux of my issue with the term "compression" in this context: Is it a smaller version of the data? Yes, the model is smaller than the total input data. But when it comes to recreating a single image, how many of the weights must be configured 'just-so' recreate an image enough to call it the same image? I'll admit ignorance here - but I also don't think that this is a thing anyone knows for sure. We can o…

Fair enough, maybe compression is a too specific term to apply here but I does not matter if it's compression or not to violate copyright. Compression was a good example to mention because it is already familiar to laypeople and established law. The main point is that it stores some sample of the original data - and if it's more it is derived from the original data (your strokes example) and applying some algorithm to reconstruct it to some approximation that we humans might find indistinguishable
Post reply on HN