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

$10M AI Mathematical Olympiad Prize

aimoprize.com

Post reply on HN
Print view View original

171–180 of 231 posts

Re: $10M AI Mathematical Olympiad Prize

#171

Post by tomatoadventure »

gregdoesit wrote: post #2

I was recently in Palo Alto, and bumped into a newly founded startup (I don't remember the name unfortunately) who set themselves the grand the vision of exactly this: winning a gold medal on the international Olympiad using AI. Their plan was to build mostly on LLMs as a start, and iterate as they go. In their barebones office space, they had a poster with a countdown of the number of weeks till the event: it was 36…

That sounds really cool, but I am having trouble finding any information about them to get in touch :(

Re: $10M AI Mathematical Olympiad Prize

#172

Post by rcpt »

auggierose wrote: post #26
gregdoesit wrote: post #2

I was recently in Palo Alto, and bumped into a newly founded startup (I don't remember the name unfortunately) who set themselves the grand the vision of exactly this: winning a gold medal on the international Olympiad using AI. Their plan was to build mostly on LLMs as a start, and iterate as they go. In their barebones office space, they had a poster with a countdown of the number of weeks till the event: it was 36…

I don't think anything came out of the Netflix prize, did it?

Netflix got enormous amounts of R and D for dirt cheap

Re: $10M AI Mathematical Olympiad Prize

#173

Post by rcpt »

jacobr1 wrote: post #142
anonylizard wrote: post #30

Earlier quoted context omitted.

I don't think the actual winning algorithm itself was used, because real world systems have more constraints/requirements than what the recommender was trained on. But that was in 2009, pre deep-learning/AI summer, and $1 mil clearly helped stimulate interest in that area. Today we see multiple billion dollar recommender systems, like Tiktok. Netflix ironically benefits the least from recommenders due to the nature o…

My understanding was that their research on what drove engagement shifted quite a bit. Things like social proof, and product patterns like auto-loading the next episode to binge drove engagement metrics. Recently there were some articles about their team custom-identifying which cuts of a video to show as a trailer maximized engagement on a personal level. In some sense that is a recommendation, but it is a broader p…

Social proof on Netflix?

Re: $10M AI Mathematical Olympiad Prize

#174

Post by Davidzheng »

auggierose wrote: post #168
Davidzheng wrote: post #81

Earlier quoted context omitted.

I think it's a very unfair assessment to make from these two examples. I think it's like saying Silicon Valley has been shutting down for a while and providing two examples of startups. I will say from my personal impression as someone in pure math, I disagree with this statement.

I don't know what "fair" has to do with that. These are two examples that caught my attention, and I might be wrong that this constitutes a trend. If you have data that says otherwise, please share.

I think the funding in pure math is not going down in recent years. Nor are math departments much smaller than they were say ten years ago. Otherwise I'm sure how best to give evidence of math departments not shutting down! Do I list healthy math departments?

Re: $10M AI Mathematical Olympiad Prize

#175

Post by dxbydt »

andyjohnson0 wrote: post #15

As the parent of a young adult currently half way through their maths undergrad, this kind of fills me with foreboding. I know that proof assistants etc have existed for quite a while now, but what with this and the murmours about OAI's Q* model, I do wonder what will happen to maths as a human endeavour - and as a enabling skill for jobs that can financially support people like my child.

Perhaps this might cheer you up - https://statmodeling.stat.columbia.edu/2015/03/17/1980-math-...

Or perhaps not. In that blog, Gelman speaks of Gregg, who ended up as a GS VP, and says -

math olympiad = high school basketball star

pro mathematician = NBA player

Goldman Sachs VP = sports hustler

I actually worked with Gregg in fixed income at that time :) Gelman's blogpost received sufficient notoriety, atleast within GS & the IB community.

Re: $10M AI Mathematical Olympiad Prize

#176

Post by logicchains »

Yhippa wrote: post #68

I'm asking this question out of ignorance: if you were able to do this, why would you make it public for $10MM instead of keeping it private and exploiting it. Say, in algorithmic trading models?

>Say, in algorithmic trading models

Higher level math is nothing like the math used in trading algorithms. It wouldn't be any more useful than a top tier PHD graduate.

Re: $10M AI Mathematical Olympiad Prize

#177

Post by oglop »

That’s, interesting I think. I’m not sure wha true domain of problems are in that arena. I’ve found great success using LLMs for teaching abstract algebra but have noticed when I switch to analysis or topology things get more wonky.

My advice, and I have zero understanding or care why the above happens, is to lean into algebra and train less on analysis. You’ll get there faster it seems.

Re: $10M AI Mathematical Olympiad Prize

#178

Post by daxfohl »

killerstorm wrote: post #122
daxfohl wrote: post #101

I'd still call this fancy autocomplete. I imagine the jump from this, to "come up with an interesting new branch of math to explore", is a long way off.

By that definition 99.99% people are also fancy autocomplete. How many people you know who came up with an interesting new branch of math?

Yeah I don't disagree with this. But such people do exist. My expectation is that IMO gold is a couple years off. But inventing new math and explaining why it's interesting and useful is much much farther away.

I can't decide whether I hope I'm wrong or not.

Re: $10M AI Mathematical Olympiad Prize

#179

Post by godelski »

anonylizard wrote: post #34
dimask wrote: post #29

Earlier quoted context omitted.

Their current ambition is to be able to solve school math, which is quite far away from solving unsolved conjectures or math olympiads. I really doubt that any of this is within LLM/transformer scope, except maybe in some auxiliary sense to other, much different architectures.

Art isn't an easier problem than math. An artbot would have sounded more sci-fi than a mathbot only 2 years ago. Yet it only took the AI world 1.5 years to go from drawing child scribbles to replicating top artists with like 90% similarity (I can barely tell the difference between AI and human drawn art anymore with the new NovelAI model). It won't be long before AI starts to go superhuman in art skills. It won't tak…

> Art isn't an easier problem than math.

In a way it is, in a way it isn't. You have to remember what is easy for machine isn't going to correlate to what is easy for us humans. Look at AI art. Closely. No, closer than that. All the detail is fucked up. Not just the hands, but the tiniest of things. Strokes, lighting, reflections, and consistency, and all that. But can I turn my friend into a convincing werewolf? Yes. Can I turn my cat into a human or Wonder Woman? No. The system isn't a "fancy copier" but it is a compression algorithm and the aforementioned tasks were only possible because lots of work training LoRAs, textual inversions, control nets, and so on (you could seriously improve GANs, VAEs, hell, even Boltzman Machines could probably do pretty well were any of these given the same research investment that diffusion has received. GANs come close but nuances like GANs having a magnitude fewer parameters).

But let's look at math, can I consistently add numbers? No. The problem is that in math, all those tiny intricate details matter. Not only that, they matter at every single step. The thing here is that these are still pattern recognition machines. But they aren't generalized machines. You can't really derive out all of math from probability distributions (or at least cleanly, but still not convinced you can). The thing is that for math to work in AI we have to address the elephants in the room: math. Yeah, math. ML people don't like it. But we gotta address the axioms in the room that we're operating under. How do we move on from machines operating on manifolds? How do we make it so data are not distributional? How do we move away from a number of unmentioned axioms remains a large open problem in AI research. One that does not get anywhere serious enough of a conversation, especially within the community. Sure, maybe transformer circuits can learn some addition by learning how to do FFTs and add in the FFT space, but you're not going to get to Abstract Algebra that way. Ideally the AI can solve problems that have no algorithms, pun intended.

Re: $10M AI Mathematical Olympiad Prize

#180

Post by auntienomen »

antoinexp wrote: post #85
anonylizard wrote: post #17

Earlier quoted context omitted.

Well math solving is exactly what the rumored Q* is aiming towards too. I don't think it'll take more than 2 years before some LLM + RL system can take the gold medal. I think companies like OpenAI are aiming for something far more ambitious, like solving a millennium prize problem (even with human assistance). That's the kind of news release that'll add another $100 billion to your market cap.

Might depend on the terms you have in mind, but current consensus seems to be more like 4-5 years as we speak on https://www.metaculus.com/questions/6728/ai-wins-imo-gold-me...

That's 4-5 years for solving Olympiad problems. Those are just very tricky high school math problems. They have solutions and can generally be solved by applying some combination of standard tricks. It's very much the sort of thing an LLM should be good at.

Solving Millennium problems is a whole different ballgame. It's not known if these problems are solvable within ZFC axioms. (In one case, the Yang-Mills prize, stating the problem mathematically is part of the challenge.) All of the obvious applications of known tricks have been tried and failed. To solve such problems, one probably has to invent new and surprising mathematical definitions, building a framework in which the problem becomes solvable. This is something that LLMs will be crap at; the process of invention is not represented in any training data we have access to.

Post reply on HN