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

Neural Network Diffusion

arxiv.org

Post reply on HN
Print view View original

61–70 of 92 posts

Re: Neural Network Diffusion

#61

Post by stale2002 »

astrange wrote: post #34
killerstorm wrote: post #23

Earlier quoted context omitted.

> which is just a belief that magic is real Is there a law of thermodynamics which prevents AI from writing code which would train a better AI? Never learned that one in school. And FYI here's OpenAI plan to align superintelligence: "Our goal is to build a roughly human-level automated alignment researcher. We can then use vast amounts of compute to scale our efforts, and iteratively align superintelligence." I guess…

> Is there a law of thermodynamics which prevents AI from writing code which would train a better AI? You need to apply Wittgenstein here. This appears to be true because you haven't defined "better". If you define it, it'll become obvious that this is either false or true, but if it is true it'll be obvious in a way that doesn't make it sound interesting anymore. (For one thing our current "AI" don't come from "writ…

> If you define it, it'll become obvious that this is either false or true

Ok. So then I guess it isn't "just a belief that magic".

Instead, it is so true and possible that you think it is actually obvious!

I'm glad you got convinced in a singular post that recursive self improvement, in the obvious way, is so true and real that it is obviously true and not magic.

Re: Neural Network Diffusion

#62

Post by rdedev »

lucubratory wrote: post #45
rdedev wrote: post #31

Earlier quoted context omitted.

Even if recursive self improvement does work out my hunch is that is going to be logarithmic instead of exponential mostly down to just availability of data. It might go beyond human intelligence but I don't think it will reach singularity

This is why the big bet for AI-assisted AI-development long term is synthetic data. A big part of the reason so much money and resources is going into synthetic data right now is not just out of economic necessity, but because there have been extremely encouraging results with synthetic data (e.g. 'Textbooks Are All You Need', AlphaZero).

I wouldn't count aplha zero since it's reinforcement learning. That technique you can generate high quality data all the time since the rules are fixed. Not everything can be trained using that way

Re: Neural Network Diffusion

#63

Post by recursivecaveat »

astrange wrote: post #53
alanbernstein wrote: post #41

Earlier quoted context omitted.

> they just come from training bigger models on the same data Are you arguing that all AI models are using the same network structure? This is only true in the most narrow sense, looking at models that are strictly improvements over previous generation models. It ignores the entire field of research that works by developing new models with new structures, or combining ideas from multiple previous works.

I sure am ignoring that, because the bitter lesson of AI is usually applicable and implies that all such research will be replaced by larger generic transformer networks as time goes on. The exception is when you care about efficiency (in training or inference costs) but at the limit or if you care about "better" then you don't.

This is kindof an odd statement because the transformer is not the most generic neural net. It's the result of many levels of improvements in architecture over older designs. The bitter lesson is methods that can scale well with compute win (alpha/beta beats heuristics alone, neural network beats alpha/beta), not that the most obvious and generic approach eventually wins. Given the context-length problems with transformers I think it's fair to say they have scaling problems.

Re: Neural Network Diffusion

#64

Post by vessenes »

SubiculumCode wrote: post #56
vessenes wrote: post #18

I wasn't sure if this paper was parody on reading the abstract. It's not parody. Two things stand out to me: first is the idea of distilling these networks down into a smaller latent space, and then mucking around with that. That's interesting, and cross-sections a bunch of interesting topics like interpretability, compression, training, over- and under-.. The second is that they show the diffusion models don't just…

Perhaps doing this to generate 10 similar but different versions of a model can then be fed into mixture of experts?

Ooh that’s a good idea! Although mistral seems to have been seeded with identical copies of mistral, so maybe it doesn’t buy you much? Sounds worth trying though!

Re: Neural Network Diffusion

#65

Post by FeepingCreature »

astrange wrote: post #16
falcor84 wrote: post #4

Seems like we're getting very close to recursive self-improvement [0]. [0] https://www.lesswrong.com/tag/recursive-self-improvement

No, this is an example of an existing technique called hypernetworks. It's not "recursive self improvement", which is just a belief that magic is real and you can wish an AI into existence. In particular, this one needs too much training data, and you can't define "improvement" without knowing what to improve to.

All current LLMs are based on the premise that magic is real and you can wish intelligence into existence; it's called "scaling laws" and "emergent capabilities".

Recursive self-improvement isn't "maybe magic is real", it's "maybe the magic we already know about stays magical as we cast our spells with more mana."

Re: Neural Network Diffusion

#66

Post by killerstorm »

astrange wrote: post #34
killerstorm wrote: post #23

Earlier quoted context omitted.

> which is just a belief that magic is real Is there a law of thermodynamics which prevents AI from writing code which would train a better AI? Never learned that one in school. And FYI here's OpenAI plan to align superintelligence: "Our goal is to build a roughly human-level automated alignment researcher. We can then use vast amounts of compute to scale our efforts, and iteratively align superintelligence." I guess…

> Is there a law of thermodynamics which prevents AI from writing code which would train a better AI? You need to apply Wittgenstein here. This appears to be true because you haven't defined "better". If you define it, it'll become obvious that this is either false or true, but if it is true it'll be obvious in a way that doesn't make it sound interesting anymore. (For one thing our current "AI" don't come from "writ…

> This appears to be true because you haven't defined "better".

Better intelligence can be defined quite easily: something which is better at (1) modeling the world; (2) optimizing (i.e. solving problems).

But if that would be too general we can assume that general reasoning capability would be a good proxy for that. And "better at reasoning" is rather easy to define. Beyond general reasoning better AI might have access to wider range of specialized modeling tools, e.g. chemical, mechanical, biological modeling, etc.

> if it is true it'll be obvious in a way that doesn't make it sound interesting anymore.

Not sure what you mean. AI which is better at reasoning is definitely interesting, but also scary.

> they just come from training bigger models on the same data.

I don't think so. OpenAI refuses to tell us how they made GPT-4. I think a big part of it was preparing better, cleaner data sets. Google tells us that specifically improved Gemini's reasoning using specialized reasoning datasets. More specialized AI like AlphaGeometry use synthetic datasets.

> Yes, OpenAI was literally founded by a computer worshipping religious cult.

Practice is the sole criterion for testing the truth. If their beliefs led them to better practice then they are closer to truth than whatever shit you believe in. Also I see no evidence of OpenAI "worshipping" anything religion-like. Many people working there are just excited about possibilities.

> Humans don't have a "recursive self-improvement" ability.

Human recursive self-improvement is very slow because we cannot modify our brains' at will. Also spawning more humans takes time. And yet humans made huge amount of progress in the last 3000 years or so.

Imagine that instead of making a new adult human in 20 years you could make one in 1 minute with full control over neural structures, connections to external tools via neural links, precisely controlled knowledge & skills, etc.

Re: Neural Network Diffusion

#67

Post by killerstorm »

astrange wrote: post #53
alanbernstein wrote: post #41

Earlier quoted context omitted.

> they just come from training bigger models on the same data Are you arguing that all AI models are using the same network structure? This is only true in the most narrow sense, looking at models that are strictly improvements over previous generation models. It ignores the entire field of research that works by developing new models with new structures, or combining ideas from multiple previous works.

I sure am ignoring that, because the bitter lesson of AI is usually applicable and implies that all such research will be replaced by larger generic transformer networks as time goes on. The exception is when you care about efficiency (in training or inference costs) but at the limit or if you care about "better" then you don't.

There's a principle more powerful than the bitter lesson: GIGO.

Training to predict internet dump can only give you so much.

There's a paper called something like "learning from textbooks" where they show that a small model trained on high-quality no-nonsense dataset can beat a much bigger model at a task like Python coding.

Re: Neural Network Diffusion

#68

Post by killerstorm »

astrange wrote: post #51
rralian wrote: post #46

Earlier quoted context omitted.

New generations build onto the scientific knowledge of previous generations. It may not be fast but that sounds like recursive improvement to me. It seems reasonable for AI to accelerate this process.

I think saying all of society is doing it is plausible, but not the same thing as a single human or AI doing it. Though… still don't think it's true. Isn't "society is self improving" what they call Whig history?

AI might have multiple instances within a single computing environment, so it's more like a population than a single individual.

I.e. "You can only use the memory which you currently use" would be a weird artificial constraint not relevant in practice.

Re: Neural Network Diffusion

#69

Post by killerstorm »

astrange wrote: post #52
always2slow wrote: post #47

Earlier quoted context omitted.

>> I guess people working there believe in magic. >Yes, OpenAI was literally founded by a computer worshipping religious cult. What cult is this?

HPMOR readers who live in group home polycules in Berkeley who think they need to invent a good computer god to stop the evil computer god.

You're confusing OpenAI and MIRI.

OpenAI founders: Ilya Sutskever, Greg Brockman, Trevor Blackwell, Vicki Cheung, Andrej Karpathy, Durk Kingma, Jessica Livingston, John Schulman, Pamela Vagata, Wojciech Zaremba, Sam Altman. All of them come from software tech industry and academic research circles, not evidence of interest in HPMOR or Yud.

Re: Neural Network Diffusion

#70

Post by killerstorm »

advael wrote: post #27
killerstorm wrote: post #26

Earlier quoted context omitted.

It's not magic, though. If AI can do work of a human, it can do work of a human. It's a trivial statement, and inability to see it is a hard cope. Are you gonna to take a bet "AI won't be able to do X in 10 years" for some X which people can learn to do now? If you're unwilling to bet then you believe that AI would plausibly be able to perform any human job, including job of AI researcher.

I don't claim it's impossible, just that there isn't a clear path from what exists now to that reality, and that the explanation presented by the above commenter (and I suppose OpenAI's website) does not clarify what they think the path is

We can reason about it without knowing the path. E.g. somebody in 1950s could say "If you have enough compute you can do photorealistic quality computer graphics". If you ask them how to build a GPU they won't know. Their statement is about principal possibility.
Post reply on HN