THE COMPUTER CHESS HAS REACHED ITS CRITICAL POINT

[ArmyBridge]

I think that the next step is to obtain that a program can emulate the strategic planning, of long term, just like a GM, at the moment the programs are too strong, and beat GM because they commit less mistake´s than the humans and the humans cannot always build a position that is 100% of strategic reasons.Years ago in chessbase I read an article that referred a Russian project if I remember well, it was called Ninja, and it played some games against Junior and others engines. Against Junior Ninja scored 50%, Junior won 2 games and Ninja 2 games, Ninja played very well in closed positions, in where the tactics did not work, and totally overplayed Junior, Junior won 2 games too, but it was open position's. The way that Ninja work is based in a certain number of high quality GM game´s, I can't remember if it was about 10,000 database that Ninja used to take them like a guide, unfortunately Ninja was very weak when the position was opened, but I believe that the idea was good and allowed to emulate the abstract thought of GM, if there were a program that could switch between both ways to evaluate...

[playjunior]

Armando, that was not Ninja. It was "Kimo", and was an April 1st joke from Chessbase. They meant Kasparov, 2:2 was his match against Junior.
That one was one of the best April 1st jokes I read/saw/heard.

[FWCC]

Maybe we are approaching a different era for computer chess.The TOP chess engines are indeed stronger than humans and can even defeat Anand maybe with ease.The question should now be how can computer chess enhance mankind in any given field?Can we apply the computational aspect of computer chess to real life to advance man in science and medicine?When chess is solved by computers maybe by 2050AD we must ask ourselves how can we apply this knowledge to help mankind.Maybe by then an algorithm used to solve chess can be applied to the sciences or medicine--who nows?



Chess is my life but my life is not only chess
FWCC

[ZeroOne]

The question should now be how can computer chess enhance mankind in any given field?Can we apply the computational aspect of computer chess to real life to advance man in science and medicine?When chess is solved by computers maybe by 2050AD we must ask ourselves how can we apply this knowledge to help mankind.Maybe by then an algorithm used to solve chess can be applied to the sciences or medicine--who nows?

I think this used to be a good goal, but as computers have become more and more powerful, the brute-force approach is working well enough so that sophisticated artificial intelligence is not needed to play chess. AI is what I think would benefit other areas of computer science (and life in general) more than the brute-force pruning algorithms chess programs generally use.

So in the field of AI research I think games like Go (see the Wikipedia article about computer Go) and Arimaa are going to provide more for us than chess programs. By the way, Arimaa can even be played with regular chess equipment.

[Nelson Hernandez]

Late to the party, but not too late to put in my two cents.

Count me in the group that sees solving chess as the ultimate objective. It was done in checkers, it can be done in chess, somehow, someday, with technology the brightest futurist among us cannot possibly imagine. But it will not happen in our lifetimes. What has to be conceded is the staggering progress that has already been made both in the hardware and software fields since the first chess program was made, and the amazing follow-on progress that is already foreseeable when one reads about the features top-end computers will have five or six years from now and considers the more developed opportunities to aggregate computer power that will naturally develop over time.

Emulating human thinking is useful up to a point, but I don't see it's value if we concede that human thinking is inherently fallible. Memory, visualization, pattern recognition, whatever you want--mistakes are made, psychological and physiological factors intrude. In my view the computer's strengths are already awesome and can only get progressively better.

My conclusion: leave well enough alone. We are making tremendous progress year by year. The endgame is for our grandchildren. We have been lucky enough to live during the opening.

[GenoM]

Is Alpha0 result of a different approach to chess programming, somehow close to what I've called the new paradigma in computer chess?

[mclane]

The goal is and was to have the strongest chess entity.
And there was a discussion which way is the most successful, the intelligent or the bean counter approach.

So far we thought, as we were told over years, that coming deep in search depth is THE solution.
And that stockfish, Komodo or Houdini were the best entities.

But is this still the truth.

The games alpha zero vs. Stockfish looked very as if a strong chess entity plays a weak chess entity. As if an intelligent entity plays a stupid one.

So maybe the paradigm comes over us and we cannot even prevent it to happen.

We cannot fight against it

Because it happened.

The question is if we accept the truth.

[Michael Sherwin]

Is Alpha0 result of a different approach to chess programming, somehow close to what I've called the new paradigma in computer chess?

The paradigm changed in January of 2006 when RomiChess was winning 100 game matches against Rybka. Just like Alpha0 won a 100 game match against SF. The difference is that Alpha0 did all its training before the match and Romi did all its training during the match. Alpha0 trained on the entire chess tree while Romi trained on just the themed position of the match.

"Hey Michael, very interesting stuff, this seems like a table-based monte carlo policy evaluation. Impressive that you would independently discover such a thing on your own."

"However this is indeed a first step towards the policy evaluation used in A0. "

The paradigm indeed did change in 2006 but very view people noticed.

[corres]

[quote="GenoM"]

I think that the future degree of development must be found in the exploitation of the new main goal of the computer chess. This has to be the phase of development in which the main goal will be a no which program or engine will reach #1 spot, but which program will manage to effect the human's fashion of game, which progarm will succeed in immitation /and may be re-creation/ of human thinking at the chess board.
So, from my point of view, things returns to the beginning /Turing, Botvinnik/, but already -- enriched by the accomplishments of the old paradigma. The new capital goal, according to me, has to be that the programs will start to play chess the way human play it. This must be the new challenge to the programmers, not who's program will be the #1 in the rating lists of SSDF, CCRL or CEGT.

[/quote]

To make a machine playing chess like a man you must model the thinking of human at first. From this the science is very far now. Moreover a machine thinking like a human claims so huge processing power for what a Personal Computer never will be capable.
It is the time to acknowledge that future is for the big networks.
The technology what gives the base of AlphaZero is the technology of big networks. Because of this AlphaZero and a machine playing like a human never will run on a home machine.
Maybe at the very far future when the quantum computer will be a common home instrument it should be happen.

[lantonov]

The technology what gives the base of AlphaZero is the technology of big networks. Because of this AlphaZero and a machine playing like a human never will run on a home machine.
Maybe at the very far future when the quantum computer will be a common home instrument it should be happen.

I am a bit more optimistic about the future. We see at present that neural networks take more and more ground pretty fast. A0 actually came too late, it should've come about 6 months earlier. Meanwhile, neural networks took over fields ranging from plumbing to astrophysics. This means that very soon a computer with 1-2 CPUs and 8-64 cores will become obsolete and regular PC will come from the shop fitted with 1, 2, 4, 8 TPUs or whatever else parallel unit comes into fashion. We have such systems here and there even now.