Reply
 
LinkBack Thread Tools Display Modes
  #1   Report Post  
Old December 6th 06, 03:33 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: May 2006
Posts: 5,003
Default Human vs Machine Matches

I would like to make a proposal, and discover your interest in two forms of
Human v Machine matches - since I am sometimes approached by both sides to
investigate possibilities.

In conversations elsewhere [Rybka thread] there are discussions of fairness,
so I propose two options;

a) Human plus database, vs Computer

b) Human vs Computers [look-ups = off] Simul.

Which of these is more interesting to the chess community, and why?

In the recent Kramnik Fritz engagement it was difficult to engage even the
chess community's interest, so both these suggestions are deliberately
different than that, since one aspect of arranging such matches is also to
obtain mainstream media interest, and whatever Kramnik/Fritz was, did not!

Is there a general sense of preference for either (a) or (b) ?

Phil Innes



  #2   Report Post  
Old December 6th 06, 04:07 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Sep 2003
Posts: 268
Default Human vs Machine Matches

In rec.games.chess.computer Chess One wrote:
I would like to make a proposal, and discover your interest in two forms of
Human v Machine matches - since I am sometimes approached by both sides to
investigate possibilities.

In conversations elsewhere [Rybka thread] there are discussions of fairness,
so I propose two options;

a) Human plus database, vs Computer


Computer only gets opening book and tablebases. Both of which humans are
capable of learning/memorizing.

Humans once had an advantage that we could prune the options down in a
reliable way. That is proving to no longer be the advantage. Had we still
have it ... I would think somebody should be arguing that humans not be able
to do this ... wait, you can't take that away from a human. Same thing with a
computer. A computer has photographic memory and thinks REALLY REALLY FAST.
That is ITS competitive advanatage. It has finally proven to be overwhelming.

What I don't get is why, when humans lose, they try to find away to invalidate
and justify their loss such that it hasn't really occurred. I guess humans
are fundamentally lawyers at heart ... all of us.

--
Thomas T. Veldhouse
Key Fingerprint: D281 77A5 63EE 82C5 5E68 00E4 7868 0ADC 4EFB 39F0


  #3   Report Post  
Old December 6th 06, 04:18 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: May 2006
Posts: 584
Default Human vs Machine Matches


Użytkownik "Chess One" napisał w wiadomo¶ci
news:[email protected]
I would like to make a proposal, and discover your interest in two forms of
Human v Machine matches - since I am sometimes approached by both sides to
investigate possibilities.

In conversations elsewhere [Rybka thread] there are discussions of
fairness, so I propose two options;

a) Human plus database, vs Computer

b) Human vs Computers [look-ups = off] Simul.

Which of these is more interesting to the chess community, and why?


I would add

c) freestyle competition humans aided with comps vs humans aided with comps.

I think it`s an interesting option that is being tested nowadays e.g. in
e-mail chess :-) but of course top-players including Topalov or Polgar also
play this kind of chess. It is more spectacular than e.g. matches Kramnik -
comps.


In the recent Kramnik Fritz engagement it was difficult to engage even the
chess community's interest, so both these suggestions are deliberately
different than that, since one aspect of arranging such matches is also to
obtain mainstream media interest, and whatever Kramnik/Fritz was, did not!

Is there a general sense of preference for either (a) or (b) ?


I think there is a general sense of preference for (c) :-)


  #4   Report Post  
Old December 6th 06, 04:30 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Jul 2006
Posts: 446
Default Human vs Machine Matches

Thomas T. Veldhouse wrote:

Computer only gets opening book and tablebases. Both of which humans are
capable of learning/memorizing.

Humans once had an advantage that we could prune the options down in a
reliable way. That is proving to no longer be the advantage. Had we still
have it ... I would think somebody should be arguing that humans not be able
to do this ... wait, you can't take that away from a human. Same thing with a
computer. A computer has photographic memory and thinks REALLY REALLY FAST.
That is ITS competitive advanatage. It has finally proven to be overwhelming.

What I don't get is why, when humans lose, they try to find away to invalidate
and justify their loss such that it hasn't really occurred. I guess humans
are fundamentally lawyers at heart ... all of us.


It can really be no surprise the computer has won - it was expected to
happen one day. There is no way I could do maths as fast as a computer
or run as fast as a car.

My computer, with a high end maths package and a mere 450 MHz CPU can
compute exactly using just integers 10000! in 32 ms (FWIW. a floating
point approximation is about 2.846259680917055 x 10^35659).

It would take me a very long time to compute 10000! and write down all
35660 digits by hand and the chances of an error would be extreamly
high. (Actually, it is quite possible it would take me longer than my
lifetime to do it.)

But I should not feel threatened. For a long time man has built tools to
do jobs faster / better than he can do. Chess can't be expected to be
any different really.

--
Dave (from the UK)

Please note my email address changes periodically to avoid spam.
It is always of the form:
Hitting reply will work for a few months only - later set it manually.

http://witm.sourceforge.net/ (Web based Mathematica front end)
  #5   Report Post  
Old December 6th 06, 05:20 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Apr 2005
Posts: 2,598
Default Human vs Machine Matches

Dave (from the UK) wrote:
My computer, with a high end maths package and a mere 450 MHz CPU can
compute exactly using just integers 10000! in 32 ms (FWIW. a floating
point approximation is about 2.846259680917055 x 10^35659).

It would take me a very long time to compute 10000! and write down
all 35660 digits by hand and the chances of an error would be
extreamly high.


We should ban computers from using fast multiplication algorithms. It
gives them an unfair advantage.


Dave.

--
David Richerby Electronic Homicidal Clock (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a clock but it wants to kill you
and it uses electricity!


  #6   Report Post  
Old December 6th 06, 06:45 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: May 2006
Posts: 5,003
Default Human vs Machine Matches

I will make a group reply.

First, I did not lead the conversation by stating what thinking there is
already on these subjects, and will try to only follow points being made, to
supplement or amplify them.

a) Human plus database, vs Computer

b) Human vs Computers [look-ups = off] Simul.

Which of these is more interesting to the chess community, and why?


Taylor Kingston wrote:

My own preference would be (b). Like Innes, I'd like to see how the
best programs perform strictly in terms of their evaluation functions,
with no pre-programmed opening encylopedia, against GM opponents. A
side benefit might be some interesting TNs.


There are two ways already considered for this to happen,

(1) that the same engine plays a variety of GM opponents. Questions to-date
are, how many GMs, how many games.

(2) a GM proposed a simul against a variety of chess engines. There can be,
say, 4 to 6, or even more. As with human simuls, when the player arrives at
the board the opponent should move, and this may require a 'command-button'
which would require the engine to play its best evaluated move. Obviously,
the faster the GM proceeds in the middle, the less time each computer has to
reach deeper-ply depths, and there is the drama!

Thomas T. Veldhouse wrote:

in preference for (a), can he say, what clock-times would be available, and
also if he would prefer many human players against the same computer engine,
or one human player against many computer engines?

Jerzy wrote:

c) freestyle competition humans aided with comps vs humans aided with comps.

I think it`s an interesting option that is being tested nowadays e.g. in
e-mail chess :-) but of course top-players including Topalov or Polgar also
play this kind of chess. It is more spectacular than e.g. matches Kramnik -
comps.

As with Mr. Veldhouse, would he too say what clock-times would be available,
and also if he would prefer many human players against the same computer
engine, or one human player against many computer engines?

Its true that S. Polgar is interested in this, and suggested some of it. I
will still not lead with 'advanced commentary', since we are still exploring
issues, but this (c) is a variant of (a) and can be considered as a seperate
topic in its own right.

Phil Innes


  #7   Report Post  
Old December 6th 06, 06:50 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Oct 2006
Posts: 1,305
Default Human vs Machine Matches

Thomas T. Veldhouse wrote:

What I don't get is why, when humans lose, they try to find away to invalidate
and justify their loss such that it hasn't really occurred. I guess humans
are fundamentally lawyers at heart ... all of us.


This is a very old, and long recognized, part of the cycle of Artificial
Intelligence research.

a) AI folk ask: what would it take to convince you that a machine is
intelligent?

b) skeptics reply: I'll believe a computer is intelligent if it can do
X. (where X changes over time - for the moment, X = win at GM level chess)

c) AI folk produce an artifact that does X

d) skeptics say: how does it work?

e) AI folk explain the mechanisms used by the artifact.

f) skeptics say: oh, well then, I was wrong. Obviously, it does not
require intelligence to do X. It only requires these cheap tricks, that
are *obviously* not related to intelligence.

g) AI folk ask: what would it take...

repeat until done

--
Kenneth Sloan
Computer and Information Sciences +1-205-932-2213
University of Alabama at Birmingham FAX +1-205-934-5473
Birmingham, AL 35294-1170
http://www.cis.uab.edu/sloan/
  #8   Report Post  
Old December 6th 06, 07:17 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: May 2006
Posts: 5,003
Default Human vs Machine Matches

This is a cheap understanding of a Turing test - which NB! does not require
a machine, and as far as AI goes is utter nonsense. What Dr. Sloan is
addressing is 'apprehension'.

It is very common [I read it 5 times here today] to hear that the human side
of things is being engaged, while from people's writing, it is clear that
this is not true.

If the term intelligence has to be used, then let it at least contain the
sense that the agency of the intelligence is capable of learning from
experience, otherwise the agency is what is called in psychology, dumber'n a
brick.

This is not a general thread to correct programmer's understanding of
themselves! lol
It proposes various agencies of human machine interaction which resolve
something of what we have not yet established, and such diversions into
rhetoric and assertion are hardly scientific, from either the prospective of
psychology or machine behavior, and in fact avoid the tests which would, if
not resolve, then clarify certain factors.

Phil Innes

"Kenneth Sloan" wrote in message
...
Thomas T. Veldhouse wrote:

What I don't get is why, when humans lose, they try to find away to
invalidate
and justify their loss such that it hasn't really occurred. I guess
humans
are fundamentally lawyers at heart ... all of us.


This is a very old, and long recognized, part of the cycle of Artificial
Intelligence research.

a) AI folk ask: what would it take to convince you that a machine is
intelligent?

b) skeptics reply: I'll believe a computer is intelligent if it can do X.
(where X changes over time - for the moment, X = win at GM level chess)

c) AI folk produce an artifact that does X

d) skeptics say: how does it work?

e) AI folk explain the mechanisms used by the artifact.

f) skeptics say: oh, well then, I was wrong. Obviously, it does not
require intelligence to do X. It only requires these cheap tricks, that
are *obviously* not related to intelligence.

g) AI folk ask: what would it take...

repeat until done

--
Kenneth Sloan
Computer and Information Sciences +1-205-932-2213
University of Alabama at Birmingham FAX +1-205-934-5473
Birmingham, AL 35294-1170
http://www.cis.uab.edu/sloan/



  #9   Report Post  
Old December 6th 06, 07:32 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Sep 2003
Posts: 268
Default Human vs Machine Matches

In rec.games.chess.computer Chess One wrote:
Thomas T. Veldhouse wrote:

in preference for (a), can he say, what clock-times would be available, and
also if he would prefer many human players against the same computer engine,
or one human player against many computer engines?


Even up. A computer opponent should be treated like any other. Both sides
treated equally.

As with Mr. Veldhouse, would he too say what clock-times would be available,
and also if he would prefer many human players against the same computer
engine, or one human player against many computer engines?


One human against one computer. Rather, it boils down to two antonists ...
one being human the other being a computer. It is really irrelavent whether
your opponent is a computer or a human, you can either beat it/him/her or not.

Its true that S. Polgar is interested in this, and suggested some of it. I
will still not lead with 'advanced commentary', since we are still exploring
issues, but this (c) is a variant of (a) and can be considered as a seperate
topic in its own right.


Crippling an opponent so the other has a chance to win is called a "handicap".
If a grand master prefers to win by handicap, so be it.

--
Thomas T. Veldhouse
Key Fingerprint: D281 77A5 63EE 82C5 5E68 00E4 7868 0ADC 4EFB 39F0


  #10   Report Post  
Old December 6th 06, 08:37 PM posted to rec.games.chess.misc,rec.games.chess.computer
external usenet poster
 
First recorded activity by ChessBanter: Nov 2006
Posts: 7
Default Human vs Machine Matches


Dave (from the UK) wrote:
Thomas T. Veldhouse wrote:

Computer only gets opening book and tablebases. Both of which humans are
capable of learning/memorizing.

Humans once had an advantage that we could prune the options down in a
reliable way. That is proving to no longer be the advantage. Had we still
have it ... I would think somebody should be arguing that humans not be able
to do this ... wait, you can't take that away from a human. Same thing with a
computer. A computer has photographic memory and thinks REALLY REALLY FAST.
That is ITS competitive advanatage. It has finally proven to be overwhelming.


Computers don't think at all. They only do calculations based on
algorithms.


What I don't get is why, when humans lose, they try to find away to invalidate
and justify their loss such that it hasn't really occurred. I guess humans
are fundamentally lawyers at heart ... all of us.


Speak for yourself. I try to understand why I lost so that I don't
make the same mistake again.


It can really be no surprise the computer has won - it was expected to
happen one day. There is no way I could do maths as fast as a computer
or run as fast as a car.

My computer, with a high end maths package and a mere 450 MHz CPU can
compute exactly using just integers 10000! in 32 ms (FWIW. a floating
point approximation is about 2.846259680917055 x 10^35659).

It would take me a very long time to compute 10000! and write down all
35660 digits by hand and the chances of an error would be extreamly
high. (Actually, it is quite possible it would take me longer than my
lifetime to do it.)

But I should not feel threatened. For a long time man has built tools to
do jobs faster / better than he can do. Chess can't be expected to be
any different really.



It is apples and oranges. The analysis required to calculate chess
variations is an order of magnitude problem. The permutations are
beyond the power of today's computers to examine all the possibilities.
Consequently, there has to be some "pruning of the branches". This is
done by evaluating each position to determine if the position is
favorable or not. This is the crux of the problem for programmers.

On another note, computers that are "booked up" should be easy for
players to play against. You don't have to worry about surprise moves.
Any innovation should take a machine by surprise. Also, if you have a
copy of Fritz at home.....

Jimmy Boy

Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Zed Rob Mitchell rec.games.chess.analysis (Chess Analysis) 97 September 2nd 06 06:19 PM
Zed Rob Mitchell rec.games.chess.politics (Chess Politics) 67 September 2nd 06 06:19 PM
Zed Rob Mitchell rec.games.chess.misc (Chess General) 99 September 2nd 06 06:19 PM
The Innes Pledge - Revised and Updated The Historian rec.games.chess.politics (Chess Politics) 53 August 31st 06 11:24 PM
For Chess Levity: Talk To Zed Rob rec.games.chess.misc (Chess General) 33 March 4th 06 04:23 PM


All times are GMT +1. The time now is 04:11 PM.

Powered by vBulletin® Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Copyright ©2004-2019 ChessBanter.
The comments are property of their posters.
 

About Us

"It's about Chess"

 

Copyright © 2017