Why don't machine learning research agents overfit?
Comments
diddid
srean
I think you should get less annoyed.
> It’s not that the simplest is more likely to be correct, it’s that you should prefer it, because it’s simple.
I don't know what Occam meant, but if you accept the formalism of PAC learning, it is more likely to be correct
https://web.archive.org/web/20170428225156/http://www.cse.bu...
https://web.archive.org/web/20130412062821/http://cs.ecs.bay...
miltava
And I like the bayesian interpretation too. Murphy's "Probabilistic machine learning" has an occam's razor section.
The idea is that a complex model explains many more configurations (datasets) than a simple one. So its (prior) probability distribution is lower on the data seen (to compensate for the other possibilities it might explain). So the (marginal) likelihood that the simple model is correct is higher if it fits the data well enough.
srean
True.
It so happens that one gets the best generalization error bounds when one combines PAC with Bayesian ideas -- the PAC-Bayesian bounds.
yorwba
The notion of "simplicity" can be completely arbitrary, though. It's enough that there are only finitely many hypotheses simpler than the best hypothesis (assuming there's such a thing as a best hypothesis). So as you eliminate hypotheses incompatible with the data, at some point you'll have eliminated all simpler hypotheses, and the simplest hypothesis left will also be the best hypothesis. If simpler hypotheses are also more likely to be correct, you get there faster, but it's not required.
srean
Except for the fact that eventually we are all dead. So it is kind of important to get there faster.
For complicated hypotheses, where complicated is defined appropriately, it takes many many examples to realize that it was a wrong hypothesis all along. There lies the rub.
For a particular instance of a learning problem we can't tell much, however using a Occams razor over many instances, one would be correct more often than not. Provided, of course, the PAC assumptions are true or they are not very far from being true. How far is not very far ? That gets very hairy to quantify.
Retric
It does get you there faster. The overwhelming majority of incorrect models are more complicated than necessary.
Removing them at effectively zero cost is extremely valuable.
zmgsabst
You choice of basis matters, eg, wavelet versus sinusoid.
ruszki
None of your links work for me.
srean
Ah! from my very dated and messy bibtex file comments. Wait, let me search for them on archive.org.
Fixed.
p-e-w
There are also various metaphysical theories that posit that the universe is algorithmically generated in some sense or the other, and from many of those theories it follows that simplicity is a fundamental feature of reality, which yields an even stronger version of Occam’s Razor.
beckhamc
And sadly, in academia, complexity (opposite of Occam's razor) is what gets you published.
captainbland
It's partly because of disagreement about what "complexity" and "simplicity" actually are. Many simple statements are in fact backed by massively complex, unstated assumptions. In attempting to deal with assumptions, scientists necessarily end up having to deal with the complexity involved in that. But the problem is if you don't engage with that, how do you know what is really more conformant with Occam's razor, as opposed to just satisfying what is readily expressed in common language?
Likewise in software development a C developer will say abstractions are not simple, a java programmer will argue that dealing with low levels details is not simple. They're both kinda right but will resort to framings which back their world view.
sillyfluke
>It’s just like the Hopper quote.
Not sure about Hopper, as I recall biographers of Lawrence of Arabia certainly made it seem like he was using the fog of war to do things he knew his superiors may object to.
Regardless, even if its misinterpreted it still has a kernal of truth and separate utility than your version, that is: the people in the field closest to the action have an operational awareness that may result in better decisions in times of urgency.
gowld
That's not true. It's pretty clear that she meant "do something you knew they were going to say no to and now you are trying to get away with something."
https://youtu.be/wHdHCoeUbU4?t=861s
> So I want to tell something to all the young people here on many many occasions you'll find it is much easier to apologize than it is to get permission. You do it then when somebody comes after you and say are you supposed to do that, "oh gee I didn't know I wasn't supposed to do that" ... so just remember it's frequently much easier to apologize than it is to get permission do it
She goes on further, explaining how to deceive your superiors to manipulate them to get what you want.
diddid
But I still don’t think that means eat all the cookies in the cookie jar and then apologize after because nobody would have given permission. That’s still about doing what you believe to be right. She even frames the fallout as “where you supposed to do that?” and not “you shouldn’t have done that”.
neutronicus
Damn no wonder she got a supercomputer named after her
bnmik2
Marcus Hutter formalized this in his AIXI work.
demibabs
Even tech giants are putting out articles seemingly fully written by Claude.
ks2048
The animated graphic labeled "Occam's razor, formalized" is bizarre. Is that really visualizing "Occam's razor, formalized"?
mrbungie
Ah, over-the-top larger-than-life LLM-isms, they are really funny when you see them in a company blog, but they are vomitive when it's your coworker copy-pasting it and insisting you on reading it.
smashah
I was expecting Occam wearing a suit.
percentcer
Nobody wants to work anymore!
serial_dev
Time to first detected slop in this article is <1s. Claudisms per paragraph is also very high.
Is it too much to ask from people to read their own article anymore?
If anyone read this at all, they would have had the ick, and would have fired off a prompt to get rid of the most popular AI slop tells...
cj
What I really dislike is having to edit my own non-LLM assisted writing to make sure I'm not accidentally confused with AI.
I caught myself writing "And that matters because..." in a HN comment but had to edit myself. Also miss uising emdashes.
bee_rider
These models are trained on human language, which belongs to us, we shouldn’t surrender it to them. Keep the em-dashes. IMO don’t overuse negative parallelisms though, they were always bad and lazy.
srean
An arms race on style would be interesting. Essentially a real life GAN.
signalbright
> Why don't machine learning research agents overfit?
they do.
sigbottle
Compression in this modern day and age is so slop.
Yes, I'm familiar with keystone results such as Solomonoff induction. It's a direct counterexample to compression - your intensional algorithm can completely outrun reality. I can literally specify a huge mega-algorithm that just searches over all possible Turing machines and evaluates them, and it's an optimal compressor. It's completely vacuous though. You can always hide the "heavy work" in your mappings and descriptions. It's ironic that a kolomogorov complexity minimizer is so loaded that it's vacuous.
This is pretty much why I roll my eyes at this point at all the compression is intelligence memes.
I wonder when intervention and causality will hit the mainstream. These tools were designed specifically to counteract purely predictive theories. But your average compression dude will hold tight to their paradigms and slogans, not realize their internal contradictions (that their own field has brought up), and then whenever a new paradigm suddenly becomes visible and mainstream, they'll latch onto that. It's not principled at all.
And to be clear - I do think intelligence is some amount of compression, and I am well aware of formal results such as the arithmetic decoding theoretical and empricial result. Just annoyed. It's literally no different than the whole Bayesianism meme. If you're not actually practicing that type of intelligence as a basis, then you don't get to go around beating the drum about how it's the ultimate reality. You're just spouting dogma to feel like part of an in-group.
nyeah
They tend not to overfit ... when there are way more data points than parameters.
wmedrano
Or when there are way more parameters than data points
gwern
The methodology is partially based on https://www.offconvex.org/2021/04/07/ripvanwinkle/ , for those thinking this sounded familiar.
32df179
Wherein Claude gives an honest assessment that it genuinely does not overfit. I also had Grok telling me that it isn't quantized.
Do the submitters really not notice that this is AI slop? Do they like this? It is a complete pain to read.
vatsachak
No point in reading anything AI related anymore. It's all slop.
We need to retvrn to rss feeds
exit
what would returning to rss feeds achieve?
nyeah
Force the clankers to support rss!
dominotw
> Machine learning, at its core, is about generalization, not memorization.
Well they memorize the patterns.
memorization doesnt mean rote learning.
red75prime
Does '110001111000000011111111111' contain n>1∧∀d(d|n→(d=1∨d=n)) as well as infinite number of other generalizations?
tomrod
I don't take issue with that. Attempting memorized pattern generalization through holdout / validation strategies is a big part of ML that you would not typically see with econometrics / psychometrics / possibly sabermetrics / most other -metrics. Philosophically the explain versus predict divide. https://www.stat.berkeley.edu/~aldous/157/Papers/shmueli.pdf
porridgeraisin
That's a bit pedantic no. Memorization in ML refers to the model having the wrong level of capacity such that it's too hard to optimise it such that it doesn't memorize the _training examples_ themselves.
I always get annoyed when people misinterpret Occam’s razor. It’s not that the simplest is more likely to be correct, it’s that you should prefer it, because it’s simple.
It’s just like the Hopper quote. She said it’s better to ask for forgiveness during the fog of war, doing something you thought was right, not to do something you knew they were going to say no to and now you are trying to get away with something.