AI language models are struggling to “get” math • Hacker News Board

Re: AI language models are struggling to “get” math

#161

Post by chaxor » Thu, Oct 13, 2022, 4:16 AM UTC

nl wrote: post #136

mavu wrote: post #114

Talking about this stuff would be so much easier if we stopped calling those software "AI". It is a machine learning algorithm. It is an electronic Parrot. thats it. And suddenly no one will wonder "OH MY WHY CANN IT NOT DO MATH< IT SMART?!?!"

> It is a machine learning algorithm. True > It is an electronic Parrot. This is incorrect, and unclear why people think this. The whole point of a good ML system is that it doesn't parrot training data. A good system can extrapolate novel answers from things it has seen. That is very far from "parroting".

Don't listen to these bots, they're all just parroting this same idea that models parrot.

Re: AI language models are struggling to “get” math

#162

Post by efishnc » Thu, Oct 13, 2022, 4:57 AM UTC

dimmuborgir wrote: post #59

yeasurebut wrote: post #55

Earlier quoted context omitted.

That’s what a musician does. They make short loops and loop them. This reads like someone who knows sheet music and theory but does not listen to music. It’s repetition of short phrases over and over. I’m not really sure what people expect of general AI trained on human generated outputs. It can’t make up anything anything “net new” only compose based upon what we feed it. I like to think AI is just showing us how si…

Those models are not trained on short loops. They are trained on whole songs just like image generation models are trained on whole images. And yet they struggle to repeat sections, modulate to a different key, create bridges, intros and outros. After a few seconds of hallucinating a melodic line they simply abandon the idea and migrate to another one. There is no global structure whatsoever.

Maybe that's the problem.

We're trying to train a full composer AI without allowing to learn about different instrument sections independently at first. The human composer will have a good idea of the different parts and know how to merge them in harmony.

I think we might get better results training separate AI systems on percussions, strings, vocals etc. then somehow create connections between them so they learn together. A band AI if you will.

We could try a BERT for each, with the generator learning to output logical sequences of sounds instead of words.

Re: AI language models are struggling to “get” math

#163

Post by fargle » Thu, Oct 13, 2022, 5:02 AM UTC

Can't drive. Bad at math. AI's are becoming almost human.

Can they do passive aggressive dickhole manager? If so, we're doomed.

Re: AI language models are struggling to “get” math

#164

Post by stevenhuang » Thu, Oct 13, 2022, 5:17 AM UTC

riskneutral wrote: post #147

1024core wrote: post #144

Earlier quoted context omitted.

See the "joke explaining" examples on Page 36 of the paper: https://arxiv.org/abs/2204.02311 and tell me if it's just "passable garbled nonsense".

That's astonishing.

Oo this reminds me. One of my favourite sci-fi novels is The Moon is a Harsh Mistress.

In it, it depicts the growth of a nascent AI from its attempts at understanding humor. The AI befriends a technician and gets the human to rate its own crafted jokes.

Eventually the AI gets really good at telling jokes, and becomes sentient as a result.

It was a very fun take on AI gaining sentience, highly recommended!

Re: AI language models are struggling to “get” math

#165

Post by stevenhuang » Thu, Oct 13, 2022, 6:07 AM UTC

Marazan wrote: post #106

woah wrote: post #84

Earlier quoted context omitted.

Have you ever used Github CoPilot? It does a lot of useful work, automating away rote typing in programming. Have you tried Dall-E or Stable Diffusion? They make good looking images. This comment seems completely unmoored from where the state of the art is right now.

Dall-E produces good looking images within certain parameters. When you are in its bounds it seems magical, once you go outside it seems like a weak joke. And many of the reasons it is bad outside its sweet spot are fundamental to how it works not a flaw that can be iterated away.

> fundamental to how it works not a flaw that can be iterated away.

Can you elaborate?

Re: AI language models are struggling to “get” math

#166

Post by stevenhuang » Thu, Oct 13, 2022, 6:12 AM UTC

mirker wrote: post #131

saghm wrote: post #117

Earlier quoted context omitted.

> How much of this is just "AI is bad at everything", but in the math case, it's easier for the lay person to tell Honestly, even as someone generally pretty dismissive of the AI hype, I'm not sure you can go that far. The whole reason we have specific mathematical notation is that human languages often are not super great at dealing with it, and English in particular is pretty abysmal for being both unambiguous and…

Because there is an algorithm for it. Convert the strings into floating point numbers, add them, convert them back to strings. It’s a leetcode medium question. It should be learnable. The article talks about abstract math questions, but even arithmetic is hard for language models.

That even arithmetic is hard for language models may suggest language models (at least in their current form) are not the most optimum for the task.

Re: AI language models are struggling to “get” math

#167

Post by 8n4vidtmkvmk » Thu, Oct 13, 2022, 6:28 AM UTC

mgraczyk wrote: post #158

smeagull wrote: post #146

Earlier quoted context omitted.

The memory is the model. Having high probability to a particular area in latent space is copying it from the memory of the training data. You're a massive pedant, and this style of comment is a large reason HN comment sections are a pain to read.

What is the difference between this and describing a human brain the same way? The brain is the model, you are "just" copying things from the memory of your brain to words that you speak or write? I don't think it's pedantic to say that an argument is wrong because it's making an incorrect claim. The claim here is that there is something different or missing between a true "model" and LLMs, and that missing thing has…

i think humans are just copying every time we give an answer without much thought. we're either regurgitating something we previously thought/solved or something we heard somewhere.

when you have to put your head down and actually think for awhile, then maybe you're doing something new or at least not within your brain-training data. I don't think AI can do this yet. it can only copy pieces of its training data out to look like something new, but it isn't really. like when i was describing a video game idea i had to a friend and he called me out on just stealing bits of other games and mashing them together. he was right. it wasn't original. and this is all the AIs can do right now.

Re: AI language models are struggling to “get” math

#168

Post by enord » Thu, Oct 13, 2022, 6:40 AM UTC

ricopags wrote: post #134

enord wrote: post #54

Earlier quoted context omitted.

Wait what? Glue as in extract high level semantic representations from _syntatic probabilities_ and pass on to appropriate domain specific tools? This is the glaring hole in LLMs, a paradoxical semantic incoherence despite impressive sentenial and gramatical coherence. As glue it is so thin as to be potable.

Quoting this tweet[0]: "Here's a brief glimpse of our INCREDIBLE near future. GPT-3 armed with a Python interpreter can · do exact math · make API requests · answer in unprecedented ways" [0] https://twitter.com/sergeykarayev/status/1569377881440276481

All that remains is a small matter of programming…

Re: AI language models are struggling to “get” math

#169

Post by nl » Thu, Oct 13, 2022, 6:42 AM UTC

chaxor wrote: post #161

nl wrote: post #136

Earlier quoted context omitted.

> It is a machine learning algorithm. True > It is an electronic Parrot. This is incorrect, and unclear why people think this. The whole point of a good ML system is that it doesn't parrot training data. A good system can extrapolate novel answers from things it has seen. That is very far from "parroting".

Don't listen to these bots, they're all just parroting this same idea that models parrot.

Not sure if you are referring to me or to the OP

Re: AI language models are struggling to “get” math

#170

Post by mjburgess » Thu, Oct 13, 2022, 7:43 AM UTC

1024core wrote: post #144

mjburgess wrote: post #27

How much of this is just "AI is bad at everything", but in the math case, it's easier for the lay person to tell . It's all just passable garbled nonesense that the reader (goes to lengths) to interept based on their prior knowledge, which is not expressed in the syntax of what these systems output. In the case of mathematics, we're far less willing to "BS away" the interpretive failures. But if we were equally deman…

See the "joke explaining" examples on Page 36 of the paper: https://arxiv.org/abs/2204.02311 and tell me if it's just "passable garbled nonsense".

> 500 bn parameter model

That solution has a compressed representation of half the internet.

NNs are "garabled nonesense" insofar as they try to generalise; insofar as they are search engines, they provide apparent sense by just repeating something in their database (= weights).

Google, likewise, returns joke explanations.