|
|
| (1) Posted by Neal Turner [Friday, Nov 7, 2014 15:40] |
It was always going to happen...
...and here we are:
http://en.chessbase.com/post/a-machine-that-composes-chess-problems
|
|
| (2) Posted by Hannu Harkola [Friday, Nov 7, 2014 16:32] |
Chess problems, perhaps, but not chess compositions.
|
|
| (3) Posted by Hauke Reddmann [Saturday, Nov 8, 2014 16:42] |
Bold, not quite convincing and smells like hype...
but it's only a start yet. If that progresses like
OTB chess, we're out of business in 20 years.
|
|
| (4) Posted by Olaf Jenkner [Saturday, Nov 8, 2014 17:22] |
The first example (Kh4/Kf4) is the ugliest problem I've ever seen. Multiple treats, mating duals ...
It is not even a correct chess problem.
|
|
| (5) Posted by Geir Sune Tallaksen Østmoe [Saturday, Nov 8, 2014 21:04] |
Currently it looks like we have nothing to fear. And the comparison to OTB chess isn't really fair. In OTB chess, you play for a mathematical result (win, draw, loss). In chess problems, you aim for artistic value. In what field have computers advanced the most - mathematics or arts?
I only briefly looked at problem 7 (Ke1 - Kh1). I counted 6 possible moves for White on move 2.
|
|
| (6) Posted by Olaf Jenkner [Sunday, Nov 9, 2014 09:33] |
Maybe it's not about chess composition but OTB play.
|
|
| (7) Posted by Hauke Reddmann [Sunday, Nov 9, 2014 19:31] |
The main letdown (if you want to term it so :-) isn't
bad key or dual or whatnot, but that the living problemist
wants to show a theme. Usually. With the AI approach, you
can only get a theme by chance, and that accidentally.
Has Iqbal tried studies yet?
|
|
| (8) Posted by Olaf Jenkner [Sunday, Nov 9, 2014 21:41] |
I assume he doesn't know anything about chess composition.
|
|
| (9) Posted by Sarah Hornecker [Sunday, Nov 9, 2014 21:46] |
Professor Iqbal has created CHESTHETICA so he knows a bit about chess composition.
The issue is, he presents a program that is in maybe the alpha stage, unable to grasp the beauty of chess composition, instead being just a cold-blooded code.
In 10 years, we will have sex robots, [1] but will they ever feel anything? Will robots one day understand beauty? Will they dream of electric sheep?
[1] http://www.extremetech.com/extreme/188047-by-2025-sexbots-will-be-commonplace-which-is-just-fine-as-well-all-be-unemployed-and-bored-thanks-to-robots-stealing-our-jobs
|
|
| (10) Posted by Olaf Jenkner [Sunday, Nov 9, 2014 22:21] |
We had them 40 years ago:
http://en.wikipedia.org/wiki/The_Stepford_Wives_%281975_film%29
|
|
| (11) Posted by Hauke Reddmann [Monday, Nov 10, 2014 17:06] |
Meh. Given that those sexbots surely are planned for male users,
a good AI-generated chess problem is the far simpler task :-)
Back to topic: In OTB AI, human+computer together are far better
than both components alone. I expect the same for AI composing
(man KS/KS helpmate, man tablemining).
|
|
| (12) Posted by Olaf Jenkner [Monday, Nov 10, 2014 18:28] |
Again: We had them 40 years ago.
Look at Torsten's problems in the PDB:
g='s#'and a='Linß'and apieces<=7
|
|
| (13) Posted by Torsten Linß [Monday, Nov 10, 2014 21:46] |
@Olaf:
An alternative search is:
(A='Sheglow' OR A='Paramonow' OR A='Paliulionis') AND (g='s#' OR g='h#') AND APIECES <= 6
@Geir:
The answer to your question "In what field have computers advanced the most - mathematics or arts?"
ist *NEITHER*. As the name says computers can compute, but they cannot do mathematics. They can run
algorithms (programs) at high speed. Therefore, they are an excellent *TOOL* for almost everything
including chess composition, mathematics and arts. A tool - not more.
|
|
| (14) Posted by Nikola Predrag [Tuesday, Nov 11, 2014 01:43] |
Computers compute but they cannot think. Perhaps they might be programmed to search for and create the original logic patterns.
On the other side, humans are more and more delighted with receiving huge amounts of data and computing, but without thinking.
Laziness might result with losing the ability of thinking, but who cares? Less thinking, more happiness :-)
|
|
| (15) Posted by Juraj Lörinc [Tuesday, Nov 11, 2014 19:32] |
Torsten and others: and do not forget Vaclav Kotesovec, who made a large amount of research in computer-helped composition. Or to much lesser extent Daniel Novomesky or myself.
|
|
| (16) Posted by Sarah Hornecker [Tuesday, Nov 11, 2014 21:12] |
I really hate how this turned into a "this composer is the worst" thread.
We should unite, but instead we are torn apart by inner conflicts...
|
|
| (17) Posted by Juraj Lörinc [Tuesday, Nov 11, 2014 23:27] |
Maybe I am totally ignorant, but where exactly do you see any inner conflict, Siegfried?
|
|
| (18) Posted by Sarah Hornecker [Wednesday, Nov 12, 2014 06:10] |
Oops, I must have misinterpreted. Sorry!
Read it again, and it is about problems with few pieces, not about being a bad composer.
|
|
No more posts |
MatPlus.Net  Forum General It was always going to happen... |
ISC 2024
