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

AI language models are struggling to “get” math

spectrum.ieee.org

Post reply on HN
Print view View original

171–180 of 201 posts

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

#171

Post by Marazan »

stevenhuang wrote: post #165
Marazan wrote: post #106

Earlier quoted context omitted.

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?

Dalle is trying to optimise a bunch of random pixels to meet the prompt.

It has a strong sense of what 1 object is. Ask it for a photo of a kitten and you get an jaw droppingly realistic photo of a kitten.

Ask it for a picture of 6 kittens and you get, well, whatever this is: https://labs.openai.com/s/PIZJe6GCfat9soN3WOV3eC9p

7 kittens, none well defined. Because the loss function it is optimising taps out once it has got "close" and close for a multiple object prompt is _a lot_ further away than for a single subject prompt (here is the 1 kitten version https://labs.openai.com/s/1aCOUxNT19kbMZZtEBG7CoFY - this is basically witchcraft it is so good, the group shot is a joke).

I suppose you could massively reduce the loss amount you are willing to accept but that doesn't guarantee dalle with optimise the correct part of the pictures - maybe I'd have just ended up with really, really good floors.

The other thing Dalle is bad at is backgrounds, and once again this is due to "optimising an error score". https://labs.openai.com/s/U1Vo2fxThuXmQZzIwLQ4g9Ai nothing about this is right. At a superficial glance it looks like the view over a city but it's a random splatching of building cutouts and when you look at the detail of the builds they are a blur of pixels that kind of approximate doors and windows but are nothing of the sort. They are super fuzzy and dream like. Because it's trying to generate an image that looks like a cityscape from its memory of Glasgow cityscapes. THere's no coherance because it's trying to covert random pixels into a cityscape not for buildings out of components that humans know go to makeup buildings.

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

#172

Post by TuringTest »

zmgsabst wrote: post #160
NateEag wrote: post #150

Earlier quoted context omitted.

> Automated theorem proving is the same problem as “complete and label the diagram”, which image generation is okay at. How so? I'm no mathematician, but I don't see how these problem types are equivalent. Could you elaborate?

Sure — the connecting topic is topos theory. For a type theory we might want to reason about, there’s a diagram (in category theory) which represents the same semantic content. These diagrams turn out to have recurring and common structures. You can represent those diagrams as adjacency matrices, where those structures have a particular “shape” in the entries. Which if you squint hard looks like an image completion p…

Who's using that topos theory, and is it well known in automated theorem proving?

I hadn't heard about it before, although I have some notions of both category theory and theorem proving.

And is your description of “complete and label the diagram” using image generation a thing that actually exists or something that has potential to be created? That could be a breakthrough in applying formal methods to real-world problems.

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

#173

Post by simplencomplex »

Why instead of expecting a language to get math, don't we use the language model to generate code, run the code and use the result?

If a language model basically is the equivalent of a dumb human, how can we expect it to be better than us at math?Most humans use calculators even for simple equations.

I've been thinking about giving access to a search engine and a command line to a GPT-3 based AI, so that it can choose to run code it wrote or to expand its knowledge, I think that's a good way to expand its capabilities, even if that's probably how we're going to get skynet in the end.

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

#174

Post by zmgsabst »

TuringTest wrote: post #172
zmgsabst wrote: post #160

Earlier quoted context omitted.

Sure — the connecting topic is topos theory. For a type theory we might want to reason about, there’s a diagram (in category theory) which represents the same semantic content. These diagrams turn out to have recurring and common structures. You can represent those diagrams as adjacency matrices, where those structures have a particular “shape” in the entries. Which if you squint hard looks like an image completion p…

Who's using that topos theory, and is it well known in automated theorem proving? I hadn't heard about it before, although I have some notions of both category theory and theorem proving. And is your description of “complete and label the diagram” using image generation a thing that actually exists or something that has potential to be created? That could be a breakthrough in applying formal methods to real-world pro…

Topos theory is used by people researching foundations, eg Michael Shulman.

> And is your description of “complete and label the diagram” using image generation a thing that actually exists or something that has potential to be created?

Somewhere between — it’s a topic being researched, but results are very early (basically, just shapes and groups).

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

#175

Post by dwringer »

p1esk wrote: post #99
dwringer wrote: post #88

Earlier quoted context omitted.

MIDI is extraordinarily expressive and is likely used to sequence a large majority of music produced within the last three decades. A lot of the instruments you hear are synthesizers or samplers running directly from MIDI. There is a lot more to what MIDI can do, and is used for, than the conception most people have from "canyon.mid" or old website background music. If an AI can do MIDI just fine then it's an extreme…

If an AI can do MIDI just fine then it's an extremely small leap to doing audio just fine. Unfortunately this is not true. It takes a huge amount of human effort to make MIDI encoded music sound good. The difference between MIDI and raw audio music generation is the same as the difference between drawing a cartoon and producing a photograph. To clarify, yes MIDI can be expressive, but what's being generated when peop…

I'm not familiar enough with existing implementations of such systems to dispute it, but there's no fundamental reason algorithmic composition systems could not include modulation parameters of all kinds (pitch/breath/effects/synthesizer controls/etc) in their output. I am envisioning a DAW set up with several VST's and samplers with routing and effects in place, then using some combination of genetic algorithms and other methods to "tweak the knobs" in the search for something pleasing.

The search space is absolutely enormous, though, so I don't dispute that it's very difficult, but I wouldn't go so far as to say that it can't be done. In such a space there are "no wrong answers" so to speak. I have a python script which creates randomized sequences of notes/rhythm and gives each one a different combination of LP/HP filters and random envelopes - it's not music but it takes on a much less mechanical quality by emulating different attacks and timbres over time, even though it's completely random.

I would go so far as to say I'd be genuinely surprised if algorithmic composition and production hasn't been used to some extent significantly greater than "basically a piano roll" in at least some of the past decade's top 40 music on the radio.

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

#176

Post by sharemywin »

enord wrote: post #54
sharemywin wrote: post #10

Earlier quoted context omitted.

The point I'm trying to make is LLMs don't need to do everything just be the glue to other systems.

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.

Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents

https://wenlong.page/language-planner/

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

#177

Post by sharemywin »

enord wrote: post #54
sharemywin wrote: post #10

Earlier quoted context omitted.

The point I'm trying to make is LLMs don't need to do everything just be the glue to other systems.

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.

you would still need an executive it would be more like a universal translator.

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

#178

Post by the_af »

gamegoblin wrote: post #130
the_af wrote: post #112

Earlier quoted context omitted.

I suspect it's not solved, because solving this (beyond some trick/toy examples) is essentially solving General AI.

It's unclear what you mean by "solved". Even a human can't turn every arbitrary problem into code to solve, but we still consider humans "generally intelligent". GPT3 can't turn as many problems into code as I can, but it can do some, and GPT4 (or whatever) will be able to do more, etc.

So is in your opinion General AI already working, only it cannot solve "as many problems"?

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

#179

Post by the_af »

nl wrote: post #133
the_af wrote: post #75

Earlier quoted context omitted.

> language models just need to translate problems into code of some kind that can be run to get the answer A huge "just"! Isn't this the magic step? Translating ambiguous symbols to meaning and combining them in meaningful ways is a big deal which, apparently, these AI models cannot do. They can just parrot things.

> Translating ambiguous symbols to meaning and combining them in meaningful ways is a big deal which, apparently, these AI models cannot do. Plenty of AI models do exactly this. Very clear examples include question answering models and code generation. In both cases novel, meaningful responses are generated. > They can just parrot things. That isn't true. While language models can parrot things it is generally specia…

So is in your opinion General AI solved? Because reliably turning symbols into meaning, outside narrow or special cases, is General AI.

In my opinion, it's not solved. GPT-3 is not General AI, it's a more clever mechanism for parroting back text it cannot truly understand. Comparisons to ways humans confuse themselves are a red herring in my opinion: the old ELIZA program could reply like a very confused or trollish human would, but nobody would argue ELIZA was a general AI.

It's just that GPT is a fascinating and more convincing illusion than ELIZA. Unlike ELIZA, it can also be used for meaningful purposes.

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

#180

Post by p1esk »

dwringer wrote: post #175
p1esk wrote: post #99

Earlier quoted context omitted.

If an AI can do MIDI just fine then it's an extremely small leap to doing audio just fine. Unfortunately this is not true. It takes a huge amount of human effort to make MIDI encoded music sound good. The difference between MIDI and raw audio music generation is the same as the difference between drawing a cartoon and producing a photograph. To clarify, yes MIDI can be expressive, but what's being generated when peop…

I'm not familiar enough with existing implementations of such systems to dispute it, but there's no fundamental reason algorithmic composition systems could not include modulation parameters of all kinds (pitch/breath/effects/synthesizer controls/etc) in their output. I am envisioning a DAW set up with several VST's and samplers with routing and effects in place, then using some combination of genetic algorithms and…

there's no fundamental reason algorithmic composition systems could not include modulation parameters of all kinds (pitch/breath/effects/synthesizer controls/etc) in their output

There is such a reason - lack of training data. Very few high quality detailed MIDI samples exist to train machine learning models like AudioLM.

For state of the art in MIDI generation, take a look at what https://aiva.ai/ produces (it's free for personal use). There you can compare raw MIDI output to an automatically generated mp3 output (using "VST's and samplers with routing and effects in place, then using some combination of genetic algorithms and other methods to "tweak the knobs" in the search for something pleasing.")

mp3 version will sound much better than raw MIDI, but (usually) significantly worse than music recorded in a studio and arranged/processed by a human.

Post reply on HN