great talk by ilya sutskever on how this https://www.youtube.com/watch?v=AKMuA_TVz3A&t=2121s
Compression is prediction
111–120 of 324 posts
Re: Compression is prediction
#112This is simply wrong.
Compression requires prediction.
The better the prediction, the better the compression, whether you are measuring fidelity or result size.
This doesn't mean that compression is prediction.
Re: Compression is prediction
#113Earlier quoted context omitted.
I wonder if the author of the article knew about the series, or do they both just independently came across this topic to talk about it.
common theory. see https://prize.hutter1.net/
And the Hutter Prize for AI which measures how good AI is by measuring how well it compresses data is over 20 years old now just to really drive the point home.
Re: Compression is prediction
#114This is the thesis behind the "Information Theory, Inference, and Learning Algorithms" course that was taught at Cambridge University. > Why unify information theory and machine learning? Because they are two sides of the same coin. In the 1960s, a single field, cybernetics, was populated by information theorists, computer scientists, and neuroscientists, all studying common problems. Information theory and machine l…
Better prediction being better compression is Shannon 1948, and the link to machine learning is MacKay 2003 at Cambridge.
Re: Compression is prediction
#115This is the thesis behind the "Information Theory, Inference, and Learning Algorithms" course that was taught at Cambridge University. > Why unify information theory and machine learning? Because they are two sides of the same coin. In the 1960s, a single field, cybernetics, was populated by information theorists, computer scientists, and neuroscientists, all studying common problems. Information theory and machine l…
I had a long ranting comment I deleted. I just don't like this trend of people presenting work in a way that makes you think some combo of 1) they discovered from scratch themselves 2) it's new 3) they didn't try to cite or acknowledge where they learned it/point to good sources 4) they don't really care about trying to teach something deeply, they want shiny stuff that makes them seem deep. This post references spec…
You're reading this the wrong way I think, citations aren't given because its obviously a pedagogical article about well established stuff. Much like you wouldn't give citations in a blog post explaining calculus.
Re: Compression is prediction
#116This is the thesis behind the "Information Theory, Inference, and Learning Algorithms" course that was taught at Cambridge University. > Why unify information theory and machine learning? Because they are two sides of the same coin. In the 1960s, a single field, cybernetics, was populated by information theorists, computer scientists, and neuroscientists, all studying common problems. Information theory and machine l…
I had a long ranting comment I deleted. I just don't like this trend of people presenting work in a way that makes you think some combo of 1) they discovered from scratch themselves 2) it's new 3) they didn't try to cite or acknowledge where they learned it/point to good sources 4) they don't really care about trying to teach something deeply, they want shiny stuff that makes them seem deep. This post references spec…
The post says this is all part of gzip and LLMs, what are you saying? I’ve been using gzip my entire life. I read between the lines “this is common knowledge” throughout the piece. Throwing in some names and dates only makes this super clear story harder to read (and more like studying then the playful exploration this post was intended as).
Re: Compression is prediction
#117Earlier quoted context omitted.
A maximally efficient compressor for the existing data distribution is not in general (and often will not be) maximally efficient for future data. The former may only be enabled by convenient local optima of the input distribution that a compressor accounting for the latter could not take advantage of. For instance, consider the distribution of strings drawn from the language '0+'. Now consider the same for the langu…
It might not be optimal, but it's not wrong to call it the best available guess. That's basically assuming Occam's razor / Solomonoff induction. Hutter published a bunch of work about what it means to have an "optimal" compressor and famously spent the past couple decades running a compression contest on the idea that it'd lead to insights in AI.
There is a distinction between a compressor for a fixed dataset and one for an unknown population from which we have a sample. The optimal compressor for the sample may be the single best guess for the population, but that's not what Solomonoff induction does. It begins with a prior that allows all possible programs, and it never assigns all probability to the single optimal compressor, so it has no problem with the all-zeroes example.
But the Hutter prize (of which I'm a big fan) is for ever-more-optimal compressors, and in fact many of the solutions don't generalise to other input data without stripping out various tricks.
Re: Compression is prediction
#118Nope; there is a bit more nuance and the distinction is important. Compression is functionally equivalent to prediction when the data distribution is exactly representative of all future problems . The story changes drastically if you want generalization -- because the test distribution could be arbitrarily different, even if it had the same support ! Eg: you observe a rare edge case in your training data and (lossy)…
The point you make is observed in compression too, though, once you start optimizing for certain use cases. When LAME encodes sound into an mp3 it doesn't give equal weight to all frequencies. It throws away data outside the range of human hearing to leave more room for what's in-range.
Re: Compression is prediction
#119Earlier quoted context omitted.
Shorter description isn't understanding, let alone of it is lossy. When you shorten a description in a lossy way, you are deciding a priori that some differences in the object don't matter, and it's not because you understand the object, but because it serves your goal of shortening the description.
Or you actually do understand it. You can't just assume smaller is better but it often is. And very often it's more information-dense.
You can compress syntax, losslessly even, with zero understanding of its semantics. Zero understanding not only imbued into the compressor/decompressor, but even the designer of the compressor doesn't require understanding the semantics. Actually, even of the syntax.
A compression program can compress a book written in a language that the author of the program doesn't understand, on a topic he knows little about.
Re: Compression is prediction
#120Not really, compression doesn't require a world model, it's mathematically pure. Any AGI system must periodically reset it's prediction since the world is inherently stochastic. When we look at prediction markets they only seem to work in the long term because human language is abstracted away from the real world, again it's mathematically pure. That's why we get bugs in code and disputes with prediction outcomes.
A better title, you can improve your compression if you make an accurate prediction. Much like how a branch predictor can make a CPU do the same work in less time. Or when your symbols are true uncompressable rules of reality (which is probably meaningless both semantically and physically again due to inherent randomness)
The main difference between minimalist and maximalists are how much that set of uncompressable rules gives you. I suspect the search space is too large. What we see in practice is that lossy rules let you cover more ground but eventually you hit a wall and have to move to a lower level of abstraction to make progress. There are 10^360 paths in a go tree, but something like 10^300,000 for molecular chemistry and that's not even all the way up (or down, say 10^3000 for the standard model of physics that's 10^900,000,000 if you want to do chemistry without chemistry abstractions.). Just semiconductor fab is 10^(10^11) so 10^(10^16) with molecular chemistry (think finding an implementation for some sort of desired self assembly outcome). AI can be way way way smarter than humans and there just not be enough energy in the universe to find these needles. So we definitely need abstractions, but those are at odds with predictions and the choice of symbols often introduces abstractions that the designer didn't consider.