[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 1758073598170786.jpg (70 KB, 640x766)
70 KB
70 KB JPG
>A file is just an array of bytes
>An array of bytes can be converted into a big number
>A big number can be expressed using a small mathematical expression
>Compress files of any size into small math expressions that take up almost no space
Why couldn't this work?
>>
That is very broadly how it works but try finding that math expression.
>>
>>107025330
Look up "pigeonhole principle".
>>
cool idea now make it happen
>>
>>107025330
>Why couldn't this work?
>A big number can be expressed using a small mathematical expression
you have no idea how time consuming of a task that is
>>
>>107025330
math expressions are also represented by arrays of bytes and therefore have their own entropy. keeping 2^10 is the same number of bytes as the number 1024. if something *looks* simpler it doesn't mean it is easier to encode. read about kolmogorov complexity
>>
>>107025330
>that take up almost no space
>Why couldn't this work?
Because you have to decompress afterwards, and that means you had to have kept track of the parts you removed or shifted around to save space. You could throw away some of that information at the expense of the quality of your decompressed output but that sucks, so the solution is buying more storage. You also have to factor in computational overhead.
>>
>>107025330
The code to do it would take up a similiar amount of space instead.
>>
File: keys.webm (1.53 MB, 800x600)
1.53 MB
1.53 MB WEBM
>>107025330
We already do this with lossy compression beyond what most people can even begin to comprehend. Modern image formats like AVIF are capable of describing rough approximations of complex images as low as 1KB, which is like a twitter post worth of information to humans.

Outside of those rough approximations you'll run into problems because things are often packed with noisy random data that can't easily be represented in any alien symbol math thing.
>>
>>107025419
that's basically every keystream based encryption
>>
File: images.png (5 KB, 298x169)
5 KB
5 KB PNG
>>107025330
This is actually the principle of generativeAI , you have a mathematical equation (Which that what a fucking AI model is)
And the model tries to represent the more accurate possible the original universe of data.
And this is the principle of why current LLM reasoning is an scam
Pic-related is an auto-encoder which is practically a neural compressor (which is what you are looking for).
>>
>>107025330
Technically reality is a very complex wave to which your brain applies Fourier transforms.

I'm not sure if that is related to your post.
>>
>>107025517
NNs can find novel patterns in data though, this has been shown long ago with image data.
>>
>>107025330

>>107025398
Irregardless of principles, I'd pidgeon her holes, if you know what I mean.
>>
>>107025923
>her
He doesn't know.
>>
>>107025330
>barrape
Get them off the machine.
>>
>>107026314
Steel wood.
>>
>>107025330
>A file is just an array of bytes
>An array of bytes can be converted into a big number
congratulations on discovering the concept of cardinality
>A big number can be expressed using a small mathematical expression
fundamentally wrong. every small mathematical expression can be mapped to a relatively small number, meaning your mapping is non-unique (you've discovered lossy compression). numbers are incompressible.
>>107025385
fucking dunning kruger retard
>>
>>107026368
You are more of a nigger retard than OP is.
>every small mathematical expression can be mapped to a relatively small number
lol

But you used the word "cardinality," so I guess that must that you're smart (lmao).
>>
>>107025682
You mean my brain collapses the wave function? My brain?
>>
>>107025330
Your plan kinda falls apart at the 'small' part of the 'small mathematical expression' part.
>>
File: .jpg (519 KB, 1536x2048)
519 KB
519 KB JPG
>>107025330
its nice that we've seen their nipples
>>
File: 1740215458744416.jpg (402 KB, 962x734)
402 KB
402 KB JPG
>>107026460
If your brain is the observer then why not?
>>
OP is onto something and retard evil niggers no like when brain is connecting dots.
Keep thinking, OP. Don't get discouraged.
>>
>>107027547
>express any number as a level expression of lower entropy
show me one example, you can't
>>
>>107025330
It's called arithmetic encoding and it was patented by IBM, chud. The patent expired like in the 2000s but it was popular anyway, mp3 uses it and jpeg too, it's much better than Huffman encoding, modern methods like ANS can be better.
>>
>>107025330
How do you compress the number 22953686867719691230002707821868552601124472329079?
>>
>>107027757
very carefully
>>
>>107025330
>A big number can be expressed using a small mathematical expression
how do you express the number 1608658603854073632126042475286356 with a small mathematical expression
>>
>>107028233
10,906916703874^2^2^2^2^2
>>
>>107025330
reminder that πfs contains every 1080p PTHC video ever made
https://github.com/philipl/pifs
>>
>>107028384
ebin.
>>
>>107025330
>mathematical expression
math is a language. as such, it would need to be compiled. into big numbers, and arrays of bytes. but anyway, apparently you can store data in wood and rocks
>>
>>107027263
Then how can i make it collapse it so its beneficial for me
>>
>>107026501
We have?
>>
>>107027757
22953686867719691230002707821868552601124472329078 + 1
>>
>>107025330
https://en.wikipedia.org/wiki/Kolmogorov_complexity
>>
>>107025330
>Why couldn't this work?
It does that's kalmagrov complexity (or however you spell his name.)
>>
>>107025330
Since Pi contains every digit combination possible, just store the indices to Pi that has your data.
>>
File: file.png (173 KB, 306x350)
173 KB
173 KB PNG
A: Kolmogorov complexity
>>
>>107029027
>btfoes compressionfags
>no response
typical
>>
>>107025330
If this us true why can't we just turn random numbers into files.
Sure, there will be a lot of garbage files but we could potentially create useful programs since every program is just a number. We just have to find the useful numbers
Imagine, fully fledged software suites created with zero coding.
>>
>>107028233
1608658603854073632126042475286355+1
youre welcome
>>
>>107025330
An compession algorithm is basically a discrete dynamic system

>read byte
>encode byte according to data model
>update data model
>read next byte

And decompression is the samy, but in reverse.
>>
>>107025330
How do you think compression works?
>>
>>107028384
how do i find the cure to cancer inside pi?
>>
>>107025330
>Why couldn't this work?
You're not the first person to think of it, the problem is called "P=NP"

https://en.wikipedia.org/wiki/P_versus_NP_problem

>>107028384
Was literally about to post this (or mention it).

This is what you (and other retards like myself) dream of doing: Instead of downloading the latest 130 gigabyte free to play game off Steam for my friends to abandon after a week, I download a message that says "using Pi starting from position X calculate until length 120 gigabytes"

Of course now I have awesome internet so it doesn't matter.
>>
>>107031238
Already posted.

>>107031284
Try it and see.

I think the problem becomes clear when you see it not so much as quantities of things (10 apples versus 20 apples) but a quantity of digits or orders of magnitude ( 10 apples or 20 apples is the same size, two digits, even though 10 is "smaller" than 20).
>>
>>107026501
>wears the cosplay panties
>not wearing actual panties underneath the cosplay panties
based

seeing the second pair of panties lazily hidden in the shot completely kills the vibe for me
>>
>>107025330
>A big number can be expressed using a small mathematical expression
Because if your big number is fully random then your "small mathematical expression" isn't gonna be any smaller.
>>
>>107025330
Isn't this basically how IEEE754 floating point encoding works?
>>
>>107025330
Fuck this gay thread, post more Dandadans.
>>
File: normal photo.png (862 KB, 640x766)
862 KB
862 KB PNG
>>107028384
Any file can be said to 'contain' any other file, as any file can be combined with an arbitrary algorithm and key combination to produce whichever other file you want. For instance, attached image can be used as key with the RelativeEmbed algorithm to 'extract' a very different image from >>107025330. Neither image contains the data on its own, but with the right algorithm either can 'contain' a different image when using the other as key.
>>
File: lol.jpg (183 KB, 1282x2015)
183 KB
183 KB JPG
>>107028384
>https://github.com/philipl/pifs
>So I've looked up my bytes in π, but how do I remember where they are?
>Well, you've obviously got to write them down somewhere; you could use a piece of paper, but remember all that storage space we saved by moving our data into π? Why don't we store our file locations there!
When you're 10 years old and your mom nags you to clean your room so you jam all the bullshit in the closet
>>
File: 9138078.png (5 KB, 181x30)
5 KB
5 KB PNG
>>107025456
The real tricks to reducing size are realizing which of the possible options can be skipped and making use of context.

For the former, the most obvious example is prime numbers. If you already know you're working only with prime numbers, the number '17' (which ordinarily would require 5 bits to represent) can be immediately derived from the number '7' (3 bits) as it's the seventh prime number. More practically, the vast majority of text uses fewer than 256 symbols, so by reserving one option for 'uncommon symbols' (which you then write out uncompressed) you can spend fewer bits on encoding the handful of symbols that are actually used in text.

For the latter, MIDI audio files are an example. Rather than storing the audio itself, they contain instructions for which instruments play which notes at which times. The parser then makes use of its soundbank to reproduce the music as instructed by the MIDI audio file. This results in multi-megabyte music sometimes being stored in as little as a few kilobytes.

The most extreme example of the latter is things like HTTP URLs or magnet links: rather than storing ANY of the actual data, you just store instructions on how to fetch the actual data from other servers on the global network you're connected to. This of course blurs the line between 'containing the data' and being unrelated data that just happens to be usable to generate other data. If a MIDI file contains music, how does a magnet link not contain the file it describes or a HTTP link not contain the page it links to?
>>
>>107025330
realistically it's easier to come up with say 500 formulas that have correct subsequences then combine them to create the entire sequence
and we call that uhhhh neural compression or something idk
>>
>>107025923
it's "unirregardless" you ESL shit
>>
>>107033264
Appreciated! I'm trying to improve every day.

Not unirregardlessness of her lack of principles, I wood steel pigeon (her) holes.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.