September 2017

S M T W T F S
     12
34 56789
10111213141516
17181920212223
24252627282930

Style Credit

Expand Cut Tags

No cut tags
Monday, October 18th, 2004 05:20 pm

Ссылки из [livejournal.com profile] ljnauka на работы по алгоритмической теории информации:

http://www.cs.auckland.ac.nz/CDMTCS/chaitin/
http://dr-gng.dp.ua/files/gnoseo.htm

Если кто имеет чего сказать, с интересом послушаю.
Monday, October 18th, 2004 04:59 pm (UTC)
а причем тут лженаука?
Monday, October 18th, 2004 06:23 pm (UTC)
Вот потому-то я и вынул это дело из той дискуссии.

Там один участник сделал хитрый ход, как это принято в других местах (в Юзнете): перекинул мостик на то, что ему интересно, и вместо продолжения разговора стал говорить об этом, попутно, правда, обложив собеседников. Но здесь не Юзнет.

Мостик он перекинул таким образом: сделал заявление, что научная теория не продукт творчества учёного, а результат более-менее автоматического изучения накопленного экспериментального материала. (Кажется, это называется индукционным подходом.) Отсюда он перешёл к не-шенноновскому определению информации как наикратчайшей программе, описывающей набор данных. Вот это последнее мне достаточно интересно, но не относится к теме [livejournal.com profile] ljnauka. Первое же относится и обсуждается (или уже не обсуждается, посмотрим) в других веточках.
(Anonymous)
Monday, October 18th, 2004 08:50 pm (UTC)
Ever heard of the Kolmogorov complexity of a source?
Monday, October 18th, 2004 11:16 pm (UTC)
Heard of it, yes. There is a grey area between math and computer science, which is either science or engineering or both, and it seems to me that Kolmogorov is on one side of it, while Chaitin is on the other one. I am more interested in the engineering aspects.
Wednesday, October 20th, 2004 09:12 am (UTC)
Kolmogorov was not a Computer Scientist. He was a mathematician, one of the last ones who made significant contributions to a broad spectrum of subfields. And yes, Kolmogorov complexity is relevant here - it also calls for shortest representation of information.
Wednesday, October 20th, 2004 09:43 am (UTC)
Well, I understand that. When I said "Kolmogorov on one side", I meant "Kolmogorov's theory on one side", a shortcut. The approach is mathematically solid, like everything Kolmogorov (the man) did, but its practical use in computer programming has been limited.

THere are interesting philosophical (or logical, if one prefers that way) issues related to this approach, as well as Chaitin's, but I am not an expert in this.
Tuesday, October 19th, 2004 06:48 pm (UTC)
Интересная и местами очень квалифицированная дискуссия на эту тему случилась 10-12 лет назад в bionet.info-theory.

Вот несколько "входных точек", но внутри там все разветвлено и длилось несколько лет:

http://bionet.hgmp.mrc.ac.uk/hypermail/bio-info/bio-info.199502/0042.html
http://bionet.hgmp.mrc.ac.uk/hypermail/bio-info/bio-info.199502/0006.html
http://bionet.hgmp.mrc.ac.uk/hypermail/bio-info/bio-info.199503/0033.html
http://bionet.hgmp.mrc.ac.uk/hypermail/bio-info/bio-info.199411/0031.html
http://bionet.hgmp.mrc.ac.uk/hypermail/bio-info/bio-info.199209/0000.html
Tuesday, October 19th, 2004 08:09 pm (UTC)
Любопытно. Спасибо!
Tuesday, October 19th, 2004 07:22 am (UTC)
Yes, I enjoy Chaitin.
Here is no "theory" at all, I find here a popular discussion about a well-known crisis in mathematics in the time of Goedel, and philosophical understanding of complicity. "Omega"-definition is a brilliant idea.
I cannot say it is very useful in "Computer Science", because a staring point of any computer problem is to estimate, is the problam solvable at all, and then to think about, how to restrict is. The modern problems of AI, like "translation", "text understanding", "image recognition" are generally solvable, but you often cannot find a "good quality solution" because of high complicity.
I do not see any "pseudo-science" here.
Tuesday, October 19th, 2004 09:55 am (UTC)
philosophical understanding of complicity

I certainly hope you meant "complexity". :-)

I do not see any "pseudo-science" here.

Neither do I, and this presents a problem of sorts. As soon as we see that something is definitely science, does it then leave the area of out interest and become offtopic?
Tuesday, October 19th, 2004 02:36 pm (UTC)
yeah, sorry, I mean "complexity" :-))
maybe this is offtopic. I cannot say that there is something false. It looks like an article in a holiday newspaper about science.
Tuesday, October 19th, 2004 08:07 pm (UTC)
No need to be sorry. ;-)

I am not yet sure about the boundaries of [livejournal.com profile] ljnauka. Here, in my blog, nothing is off topic.