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

AI for math resources, and erdosproblems.com

terrytao.wordpress.com

Post reply on HN
Print view View original

21–30 of 39 posts

Re: AI for math resources, and erdosproblems.com

#21

Post by benreesman »

kubrickslair wrote: post #20
benreesman wrote: post #6

Earlier quoted context omitted.

I’m generally a bit of an LLM skeptic, at least as concerns extravagant claims around “AGI” or self-improvement of any of that nonsense. But there are activities where an inherently high-temperature assistant (e.g. some Instruct-inspired tune) are a great fit, and those are almost definitionally at the boundary of “objectively faithful” and “a bit stochastic”. I’ve never done any work in novel mathematics myself, but…

The impact extends beyond mathematics. We have seen some pretty positive outcomes for LLMs as an assistant for scientists - with a lot of custom agent dev though - and not off-the-shelf models. This is already being used by a few hundred scientists in a lab. We are aiming to extend to thousands next year with a focus on CS, climate & bio.

I think this readership would be highly interested by anything hundreds of serious scientists are doing in serious work.

Can you share any details?

Re: AI for math resources, and erdosproblems.com

#22

Post by 77pt77 »

kevinventullo wrote: post #9
benreesman wrote: post #6

Earlier quoted context omitted.

I’m generally a bit of an LLM skeptic, at least as concerns extravagant claims around “AGI” or self-improvement of any of that nonsense. But there are activities where an inherently high-temperature assistant (e.g. some Instruct-inspired tune) are a great fit, and those are almost definitionally at the boundary of “objectively faithful” and “a bit stochastic”. I’ve never done any work in novel mathematics myself, but…

FWIW as a former mathematician, I believe that if LLM’s can crack mathematics in the sense of say, solving the Riemann Hypothesis through chains of logical reasoning, then we’d be in the back half of reaching “AGI” in the sense of solving broad classes of general problems like drug discovery, or getting fusion to work, or even improving itself.

Most likely it will just generate bullshit proofs, claim otherwise and then just flat out lie to your face.

Re: AI for math resources, and erdosproblems.com

#23

Post by kubrickslair »

benreesman wrote: post #21
kubrickslair wrote: post #20

Earlier quoted context omitted.

The impact extends beyond mathematics. We have seen some pretty positive outcomes for LLMs as an assistant for scientists - with a lot of custom agent dev though - and not off-the-shelf models. This is already being used by a few hundred scientists in a lab. We are aiming to extend to thousands next year with a focus on CS, climate & bio.

I think this readership would be highly interested by anything hundreds of serious scientists are doing in serious work. Can you share any details?

Thanks! Most of the current work is not yet out as press or blogs - we generally do that post code release/ publications/ arXiv submissions. Will share them as they are made public.

Meanwhile, this is the blog of our earliest work from January that will be published in ICML24: https://blog.allenai.org/data-driven-discovery-with-large-ge...

Also if you are interested - happy to correspond more by email. Lot more to share privately.

Re: AI for math resources, and erdosproblems.com

#24

Post by benreesman »

kubrickslair wrote: post #23
benreesman wrote: post #21

Earlier quoted context omitted.

I think this readership would be highly interested by anything hundreds of serious scientists are doing in serious work. Can you share any details?

Thanks! Most of the current work is not yet out as press or blogs - we generally do that post code release/ publications/ arXiv submissions. Will share them as they are made public. Meanwhile, this is the blog of our earliest work from January that will be published in ICML24: https://blog.allenai.org/data-driven-discovery-with-large-ge... Also if you are interested - happy to correspond more by email. Lot more to sh…

https://arxiv.org/pdf/2402.00159

MIT, the Steel Factory, Cal, UDub and what, 37 authors pushed a preprint and no one heard about it?

@dang I can’t be the only one who has had it on this GPT-5 test flight shit. I don’t care how many clean Azure IPv4 blocks Altman has, throw this motherfucker out.

Re: AI for math resources, and erdosproblems.com

#25

Post by kubrickslair »

benreesman wrote: post #24
kubrickslair wrote: post #23

Earlier quoted context omitted.

Thanks! Most of the current work is not yet out as press or blogs - we generally do that post code release/ publications/ arXiv submissions. Will share them as they are made public. Meanwhile, this is the blog of our earliest work from January that will be published in ICML24: https://blog.allenai.org/data-driven-discovery-with-large-ge... Also if you are interested - happy to correspond more by email. Lot more to sh…

https://arxiv.org/pdf/2402.00159 MIT, the Steel Factory, Cal, UDub and what, 37 authors pushed a preprint and no one heard about it? @dang I can’t be the only one who has had it on this GPT-5 test flight shit. I don’t care how many clean Azure IPv4 blocks Altman has, throw this motherfucker out.

Haha, I think it was covered at HN. But did not get as much attention as it should have.

Re: AI for math resources, and erdosproblems.com

#26

Post by pineaux »

benreesman wrote: post #15
mindcrime wrote: post #11

Earlier quoted context omitted.

I’m generally a bit of an LLM skeptic, at least as concerns extravagant claims around “AGI” or self-improvement of any of that nonsense. Of all the highly optimistic applications for 2024 LLMs, mathematics sounds pretty plausible? You probably already know this, but I think it bears pointing out explicitly: in this context we're not talking (only) about LLM's. There's a LOT more to AI than just Large Language Models,…

I’m an advocate for shifting resources from pure-scale transformer work to other promising avenues in both representation and reinforcement learning.

Never going to happen until upscaling does not return easy gains.

Re: AI for math resources, and erdosproblems.com

#27

Post by seanhunter »

iamcreasy wrote: post #7

The article or the linked Google doc doesn't mention how AI for Math resources is different than other resources list out there?

To pick two examples which show the difference other lists might have an article about "Copilot" whereas this one has "Getting copilot to work with Lean4". It's not "Code assistant" it's getting that working with coq etc.

ie specifically about AI for maths rather than AI in general

Re: AI for math resources, and erdosproblems.com

#28

Post by rldjbpin »

while i am too far removed to make wise comments, it is fascinating that applied maths is being put to use for finding new patterns.

after going through the google docs for resources, i get a sense of some high level topics. however, i am confused about the current paradigms. are ml models being used to reduce the search space or as a way to let computers bring about serendipity?

it seems like these approaches are like graphing libraries or monte carlo sims that help the user see the bigger picture and then put in their own thinking.

Re: AI for math resources, and erdosproblems.com

#29

Post by pietroppeter »

Calling out about the second announcement (erdosproblems.com) that there is a call for help and I think many people here might be able to help and some may be happy to do it. Skills required include GitHub, web design, Python/Html/Sql, …

Details: https://www.erdosproblems.com/help

Re: AI for math resources, and erdosproblems.com

#30

Post by needfish »

pietroppeter wrote: post #29

Calling out about the second announcement (erdosproblems.com) that there is a call for help and I think many people here might be able to help and some may be happy to do it. Skills required include GitHub, web design, Python/Html/Sql, … Details: https://www.erdosproblems.com/help

As a college student learning Web stuff on my own, this is a pretty nice thing to spend summer on.
Post reply on HN