Lyudmil Tsvetkov wrote:The training matches are different from the 100 games match with Stockfish.
Yes, the plot on the diagram is the training game, but 100 games per openning was played. 50-50, and the score below the diagram is on AlphaZero perspective.
12 openings with reversed colours don't square in any way with 100 played games, so did they actually left some openings played more than others, or did not they flip colours?
12 opennings x 100 = 1,200 games total.
Before we were talking about 300 and 100, now 1200 suddenly appears...
The 64/36 score certainly comes from 100 games, unless they assigned random points for a win.
And in that sample, I see Alpha playing just 1.d4 and 1.Nf3.
Read the series of posts properly. It is 100 games per openning, you clearly don't understand Table 2.
300 games because you were talking about 1.e4 earlier which appears in 6 diagrams.
How many is 50 x 6 ?
You were claiming that AlphaZero didn't play 1.e4, i told you it did! 300 times it did play 1.e4 against SF8.
See the total summation below: 1,200 games for all 12 opennings. Come on man, even this very
basic stuff we argue?
MikeGL wrote:See the total summation below: 1,200 games for all 12 opennings. Come on man, even this very
basic stuff we argue?
I suppose when the Wheel was first invented, there must have been people like Lyudmil Tsvetkov around, who said that it would never work.
Throughout the History of Mankind there have always been people who flatly refused to believe in revolutionary discoveries and inventions.
I suppose this falls in the same Category.
Since no one of us has access to TPU it's only fair to count in terms of what is available (for example if we had Alpha0 x64 binary compile and wanted to run it at home).
1TPU ~ 30xE5-2699v3 (18 cores machine).
4TPUs ~ 2000 Haswell cores
Apples and Bananas,
Stockfish is not able to make use of these TPUs,
and AlphaZero depends probably heavily on floating point operations (maybe half precision) to query the neural network.
So the question might be, if an stripped down x86-64 version of AlphaZero, with only some hundred or thousand of nps, is still able to beat Stockfish.....dunno.
--
Srdja
No original paper is comparing apples and bananas. SF is running on general purpose hardware. TPUs are not commercially available so running alpha0 on TPUs is giving it huge unfair advantage.
It would be like running SF on special hardware where search is happening on conventional CPU and all evaluation is handled with hundreds if not thousands of FPGAa, something like DeepBlue. Then we could say that comparison is fair.
Even in this setup, if it was the most recent version of Brainfish (so with opening book), and normal TC like 40/40 not 1move/min, Alpha0 would probably loose.
- i was wrong, it looks like they are doing 8 bit integer operations for inferencing
Lyudmil Tsvetkov wrote:The training matches are different from the 100 games match with Stockfish.
Yes, the plot on the diagram is the training game, but 100 games per openning was played. 50-50, and the score below the diagram is on AlphaZero perspective.
12 openings with reversed colours don't square in any way with 100 played games, so did they actually left some openings played more than others, or did not they flip colours?
12 opennings x 100 = 1,200 games total.
Before we were talking about 300 and 100, now 1200 suddenly appears...
The 64/36 score certainly comes from 100 games, unless they assigned random points for a win.
And in that sample, I see Alpha playing just 1.d4 and 1.Nf3.
Read the series of posts properly. It is 100 games per openning, you clearly don't understand Table 2.
300 games because you were talking about 1.e4 earlier which appears in 6 diagrams.
How many is 50 x 6 ?
You were claiming that AlphaZero didn't play 1.e4, i told you it did! 300 times it did play 1.e4 against SF8.
See the total summation below: 1,200 games for all 12 opennings. Come on man, even this very
basic stuff we argue?
We are talking here of the 100 game match, for which we have the pgn.
Do you have the pgn for the training games, which, btw., are claimed to run into the thousands?
MikeGL wrote:See the total summation below: 1,200 games for all 12 opennings. Come on man, even this very
basic stuff we argue?
I suppose when the Wheel was first invented, there must have been people like Lyudmil Tsvetkov around, who said that it would never work.
Throughout the History of Mankind there have always been people who flatly refused to believe in revolutionary discoveries and inventions.
I suppose this falls in the same Category.
That is the point, Anil, there is nothing revolutionary here, except the big hardware.
Deep Blue followed the same approach: big hardware.
Revolutionary discoveries require a lot of effort, while the Alpha team has put all the brunt on the hardware. That is why it can not expect any significant discoveries.
I am eagerly waiting for the time(which will never come), when they will release their single-core engine, around 2850, to see whether you are going to buy it.
Beyond obscure hypothesis about hardware, the point is the METHOD,
something totally revolutionary, the fact that the learning process took only hours, days
from zero to super GM. The games are astounding. Why try to disparage this?
To say that i'm impressed is a understatement. Shocking.
Take note that SF was at 64 cores!
Not so fast folks...And just what were the exact parameters of the test?
What opening book....? ..time control was 1 minute per move!!..
...more testing needed..A Lot More Testing.....Lets Wait and See.....Hopefully This time they won't dismantle the chess program/engine!!...AR
I totally agree. I am sure Google is onto something here but its very early. Lets examine the conditions carefully that the match was played in. Oh, what was that? They dismantled the thing? Nuts.
Wonderful result! Have they publicized the shogi results or published those match games anywhere? Unlike the chess and go results, that appeared to be a clear advance over the state-of-the-art.
I see there is lots of bickering about the fairness of the Stockfish match. I look forward to AlphaZero's participation in the World Computer Chess Championship, where all participants can use as much preparation and beefy hardware as they want! That appears to be the appropriate forum for users of custom hardware to strut their stuff. Heck, DeepMind would be superstars at the attached computer games conference!
Lyudmil Tsvetkov wrote:The training matches are different from the 100 games match with Stockfish.
Yes, the plot on the diagram is the training game, but 100 games per openning was played. 50-50, and the score below the diagram is on AlphaZero perspective.
12 openings with reversed colours don't square in any way with 100 played games, so did they actually left some openings played more than others, or did not they flip colours?
12 opennings x 100 = 1,200 games total.
Before we were talking about 300 and 100, now 1200 suddenly appears...
The 64/36 score certainly comes from 100 games, unless they assigned random points for a win.
And in that sample, I see Alpha playing just 1.d4 and 1.Nf3.
Read the series of posts properly. It is 100 games per openning, you clearly don't understand Table 2.
300 games because you were talking about 1.e4 earlier which appears in 6 diagrams.
How many is 50 x 6 ?
You were claiming that AlphaZero didn't play 1.e4, i told you it did! 300 times it did play 1.e4 against SF8.
See the total summation below: 1,200 games for all 12 opennings. Come on man, even this very
basic stuff we argue?
Do you have the pgn for the training games, which, btw., are claimed to run into the thousands?
Note that all Training games are self-play (no SF8 involved). The 1,200 are all match games against SF8.
No data given in PDF about the total number of self-play for learning, neither were the self-play PGN published. Only SF8 match was published.
100 game match versus SF8 was played on all 12 common ECO openings.
I guess you are confused of the plotted graph being put beside the
diagram. The graph is self-play, the diagram is SF8 match.
Try to read the caption of Table 2 properly. 37 times if need be.
