Live data from Hacker News

How to fit any dataset with a single parameter

arxiv.org

41–50 of 155 posts

Re: How to fit any dataset with a single parameter

#41
post #17

This reminds me of a joke idea I read somewhere: You can encode the entire Encyclopaedia Britannica using a single mark on a simple stick! Just encode the text as a ascii codes after the decimal dot of a zero. (0.656168.. etc). Then just mark that ratio of the sticks length and you're done...

That is from the novel "Hard-Boiled Wonderland and the End of the World" by Haruki Murakami[0].

[0]: https://everything2.com/title/Encyclopedia+on+a+toothpick

Re: How to fit any dataset with a single parameter

#42
post #17

This reminds me of a joke idea I read somewhere: You can encode the entire Encyclopaedia Britannica using a single mark on a simple stick! Just encode the text as a ascii codes after the decimal dot of a zero. (0.656168.. etc). Then just mark that ratio of the sticks length and you're done...

I don't see this as a joke, but a radical and important point. Reality, in being geometrical, is infinitely informationally dense (with a discrete conception of information). This distinction between geometrical space and time, and discrete algorithmic computability is unbridgeable. And hence there is an extemely firm footing on which to reject: AI, brain scanning readers, teleporters, etc and most sci-fi computation…

Existence of analog reality is quite problematic, because it opens can of worms - eg. hypercomputation is possible.

Re: How to fit any dataset with a single parameter

#43

This reminds me of the "worst way to ask a user for a telephone number" UI, after one UI forced the user to select each digit in a 0-9 dropdown. The specific one I'm thinking of is just spit out a scrollable numeric string of Pi, and make the user scroll until their phone number was the digits of Pi that matched it. My (7 digit) phone number occurs after digit position 9_000_000, and occurs 21 times in the first 200_…

I think it is still not proven that Pi contains all possible strings.

https://math.stackexchange.com/questions/216343/does-pi-cont...

Re: How to fit any dataset with a single parameter

#44
post #17

This reminds me of a joke idea I read somewhere: You can encode the entire Encyclopaedia Britannica using a single mark on a simple stick! Just encode the text as a ascii codes after the decimal dot of a zero. (0.656168.. etc). Then just mark that ratio of the sticks length and you're done...

One theoretical objection to this idea is that distances measured have an accuracy limit of Planck length (1.616e-35 meters). So if your number needs more precision then "just mark" step can't be done.

Re: How to fit any dataset with a single parameter

#45
post #17

This reminds me of a joke idea I read somewhere: You can encode the entire Encyclopaedia Britannica using a single mark on a simple stick! Just encode the text as a ascii codes after the decimal dot of a zero. (0.656168.. etc). Then just mark that ratio of the sticks length and you're done...

My shot at the calculation for fun :-)

Stick encoding with graphite resolution (0.335 * 10^-9 meter) [1]: "Uti" (31 bits -> 3 UTF8 characters)

Stick encoding with Planck resolution (1.616255 * 10^-35 meter) [2]: "Utility ought " (115 bits -> 14 UTF8 characters)

Complete first sentence: "Utility ought to be the principal intention of every publication." [3]

It appears that this storage scheme may not be suited towards the safekeeping of literature.

[1] https://www.wolframalpha.com/input/?i=floor%28floor%28log_2%...

[2] https://www.wolframalpha.com/input/?i=floor%28floor%28log_2%...

[3] https://digital.nls.uk/encyclopaedia-britannica/archive/1441...

Re: How to fit any dataset with a single parameter

#46
post #17

This reminds me of a joke idea I read somewhere: You can encode the entire Encyclopaedia Britannica using a single mark on a simple stick! Just encode the text as a ascii codes after the decimal dot of a zero. (0.656168.. etc). Then just mark that ratio of the sticks length and you're done...

I don't see this as a joke, but a radical and important point. Reality, in being geometrical, is infinitely informationally dense (with a discrete conception of information). This distinction between geometrical space and time, and discrete algorithmic computability is unbridgeable. And hence there is an extemely firm footing on which to reject: AI, brain scanning readers, teleporters, etc and most sci-fi computation…

> And hence there is an extemely firm footing on which to reject: AI

This is like saying it's impossible to build a water pump without solving the quantum mechanical interactions that govern water flow.

Re: How to fit any dataset with a single parameter

#47
post #36

Earlier quoted context omitted.

Reality isn't quite infinitely dense though, it's hard to imagine how you could encode data geometrically over a distance smaller than the planck length, for example. Reality just has a very, very fine resolution.

A quick Google shows 6.25e+34 plank lengths in a meter. So, seems impossible to measure and make the mark. Edit: As a fraction of appended ascii codes.

And if the mark were instead the removal of a single atom in a 40cm length, the resolution would be even poorer.. able to encode a single 30 bit number by my very rough calculation.

Re: How to fit any dataset with a single parameter

#49
I've thought about this for a while. I didn't go to the trouble of actually inventing the function, but just the fact that it ought to be possible in principle is sufficient to change your perspective on some things. Most notably, parameter count based information criteria.

AIC, BIC, et al. need to be reformulated for each model in which they're used, and not all parameters can be treated as fungible.

Re: How to fit any dataset with a single parameter

#50

Earlier quoted context omitted.

I don't see this as a joke, but a radical and important point. Reality, in being geometrical, is infinitely informationally dense (with a discrete conception of information). This distinction between geometrical space and time, and discrete algorithmic computability is unbridgeable. And hence there is an extemely firm footing on which to reject: AI, brain scanning readers, teleporters, etc and most sci-fi computation…

I just wanted to add that reality being a non-discrete geometric thing is still an assumption - we don't know for sure that it isn't discrete and a lot of quantum stuff points more toward a discrete reality than a continuous one. So assuming continuous space/time and discrete information then I'd agree, but as far as we know space/time aren't continuous, but just appear that way to us. It doesn't seem like we know fo…

I think for my purposes defining continuous = unmeasurably discrete produces the same results.

Ie., there is an irreducible geometrical continuity in the sense that no discontinuity can ever appear. The state density is maximal.

via this route we reporduce the same point: computationalism/simulation'ism' is then just the thesis that computers qua measurably discrete systems can realise dense unmeasurable discrete systems.

This can be shown to be impossible with much the same argument: spatial and temporal geometrical properties obtain in virtue of dense discreteness; and fail to obtain at measureable levels.

The key property of continuity is its irreducibility to measurably discrete systems. That irreducibility isn't, however , limited to continuity .

Wolfram makes this point about the failures of reductionism in a perfectly discrete context, ie., that no CA can compute a CA whose complexity is greater than it can summarise.

I prefer to press a continuous angle: our best theories of all of reality are continuous and geometrical . That energy levels are discrete in bounded quantum systems has almost nothing to do with the indispensability of contintuous mathemtatjcs in every known physical system -- including that very bounded wavefn

Post reply on HN