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

Imagen, a text-to-image diffusion model

gweb-research-imagen.appspot.com

Post reply on HN
Print view View original

561–570 of 661 posts

Re: Imagen, a text-to-image diffusion model

#561

Post by actionfromafar »

dclowd9901 wrote: post #515
qz_kb wrote: post #437

I have to wonder how much releasing these models will "poison the well" and fill the internet with AI generated images that make training an improved model difficult. After all if every 9/10 "oil painted" image online starts being from these generative models it'll become increasingly difficult to scrape the web and to learn from real world data in a variety of domains. Essentially once these things are widely availa…

Eventually the only jobs humans will have is training AI to act human. Sounds very Philip K Dick now that I think about it.

The transition will be complete when some AI can fool/bribe the other AIs that its workers are human.

Re: Imagen, a text-to-image diffusion model

#562

Post by PoignardAzur »

woeirua wrote: post #442
axg11 wrote: post #384

Earlier quoted context omitted.

I firmly believe that ~20-40% of the machine learning community will say that all ML models are dumb statistical interpolators all the way until a few years after we achieve AGI. Roughly the same groups will also claim that human intelligence is special magic that cannot be recreated using current technology. I think it’s in everyone’s benefit if we start planning for a world where a significant portion of the expert…

You should be much more concerned about the prospect of nuclear war right now than the sudden emergence of an AGI.

Is it really that simple?

We can worry about two things at once. We can be especially worried that at some point (maybe decades away, potentially years away), we'll have nuclear weapons and rampant AGI.

Re: Imagen, a text-to-image diffusion model

#564

Post by aembleton »

exac wrote: post #462
addandsubtract wrote: post #174

Earlier quoted context omitted.

I agree. How cool would it be to get an 8 min version of your favorite song? Or an instant DnB remix? Or 10 more songs in the style of your favorite album?

You can sort of do that with https://fairuseify.ml

I tried that site and the music sounds the same. I wonder if you can use this to bypass YouTube content ID check.

Re: Imagen, a text-to-image diffusion model

#565

Post by zimpenfish »

Nition wrote: post #550
joeycodes wrote: post #520

Earlier quoted context omitted.

I've started to ask myself if my own creativity is a result of random sampling from the diffusion tapestry of associated memories and experience on that topic.

I do wonder what Dall-E 2 would output for a request along the lines of "A still life of a vase of flowers in a completely new art style."

Don't have access to Dall-E 2 or Imagen but I do have [1] and [2] locally and they produced [3] with that prompt.

[1] https://github.com/nerdyrodent/VQGAN-CLIP.git [2] https://github.com/CompVis/latent-diffusion.git [3] https://imgur.com/a/dCPt35K

Re: Imagen, a text-to-image diffusion model

#566

Post by nootropicat »

planb wrote: post #491

Seeing the artificial restrictions to this model as well as to DALL-E 2, I can't help but ask myself why the porn industry isn't driving its own research. Given the size of that industry and the sheer abundance of training material, it seems just a matter of time until you can create photo realistic images of yourself with your favourite celebrity for a small fee. Is there anything I am missing? Can you only do this…

This is completely outside the competency of the current porn industry.

You gave an example of a still image, but it's going to end up with an AI generating a full video according to a detailed text prompt. The porn industry is going to be utterly destroyed.

Re: Imagen, a text-to-image diffusion model

#567

Post by abel_ »

qz_kb wrote: post #437

I have to wonder how much releasing these models will "poison the well" and fill the internet with AI generated images that make training an improved model difficult. After all if every 9/10 "oil painted" image online starts being from these generative models it'll become increasingly difficult to scrape the web and to learn from real world data in a variety of domains. Essentially once these things are widely availa…

On the contrary -- the opposite will happen. There's a decent body of research showing that just by training foundation models on their outputs, you amplify their capabilities.

Less common opinion: this is also how you end up with models that understand the concept of themselves, which has high economic value.

Even less common opinion: that's really dangerous.

Re: Imagen, a text-to-image diffusion model

#568

Post by dave_sullivan »

geonic wrote: post #560

Can anybody give me short high-level explanation how the model achieves these results? I'm especially interested in the image synthesis, not the language parsing. For example, what kind of source images are used for the snake made of corn[0]? It's baffling to me how the corn is mapped to the snake body. [0] https://gweb-research-imagen.appspot.com/main_gallery_images...

Well, first they parse the language into a high level vector representation. Then they take images and add noise and train a model to remove the noise so it can start with a noisy image and produce a clear image from it. Then they train a model to map from the word representation for text to the noisy image representation for the corresponding image. Then they upsample twice to get to good resolution.

So text -> text representation -> most likely noised image space -> iteratively reduce noise N times -> upsample result

Something like that, please correct anything I'm missing.

Re: the snake corn question, it is mapping the "concept" of corn to the concept of a body as represented by intermediary learned vector representations.

Re: Imagen, a text-to-image diffusion model

#569

Post by dougmwne »

sydthrowaway wrote: post #391
dougmwne wrote: post #283

Earlier quoted context omitted.

I think the serious answer is that it is yet another labor multiplier like electricity and software. Our tech since the industrial revolution has allowed us to elevate ourselves from a largely agrarian society to space and cyberspace. AI, by all appearances, continues to be a tool, just the latest in a long line of better tools. It still requires a human to provide intent and direction. Right now in my job, I command…

Well, anyone over 40 will be fucked. There goes your utopia.

Computers didn't fuck anyone over 40, but they did create new opportunities for young people that slowly took over the labor market and provided a steady stream of productivity growth. Right now these are impressive benchmarks and neat toys that cost millions to train. This is going to be a slow transition to a new paradigm. We are not going to end up in a utopia any more than computers created a utopia.

Re: Imagen, a text-to-image diffusion model

#570

Post by dougmwne »

slimsag wrote: post #465
rhacker wrote: post #448

Earlier quoted context omitted.

Look at carpentry blogs, recipe blogs. Nearly all of it is junk content. I bet if you combined GPT and imagen or dalle2 you could replace all of them. Just provide a betty crocker recipe and let it generate a blog that has weekly updates and even a bunch of images - "happy family enjoying pancakes together" I can see the future as being devoid of any humanity.

I wrote a comedic "Best Apache Chef recipe" article[1] mocking these sites. I guess the concern would be: If one of these recipe websites _was_ generated by an AI, the ingredients _look_ correct to an AI but are otherwise wrong - then what do you do? Baking soda swapped with baking powder. Tablespoons instead of teaspoons. Add 2tbsp of flower to the caramel macchiato. Whoops! Meant sugar. [0] http://slimsag.com/best-…

Then we will still need humans in the loop to do the cherry picking/supervised learning. Gibberish recipes need to be flagged and interesting new creations need to be promoted. The input can be fed back into the model till the model contains accurate representations of the chemical reactions of cooking ingredients and the neuronal wiring of the human olfactory system.
Post reply on HN