Chess player 'won't play for Iran' due to ban on Israeli players
Iran’s top rated chess champion
would be the country's second sports figure in recent months to renounce
his citizenship over pressures on Iranian athletes to forego matches
with Israeli competitors.
DUBAI - Iran’s top rated chess champion has decided not to play
for his country, Iranian news agencies reported on Tuesday, in an
apparent reaction to Tehran’s informal ban on competing against Israeli players.Alireza Firouzja,
the world’s second-highest rated junior player, would be the second
Iranian sports figure in recent months to try to renounce his
citizenship over pressures on Iranian athletes to forego matches with
Israeli competitors.
In October, Iran was banned indefinitely from
international judo by the sport’s world body until it could guarantee
that its athletes would be allowed to face Israelis. The move came after
an Iranian judoka said he was pressured to drop out of bouts to avoid
facing an Israeli athlete.“Firouzja
has made his decision and has told us that he wants to change his
nationality,” the president of Iran’s Chess Federation, Mehrdad
Pahlavanzadeh, told the semi-official news agency Tasnim.“Firouzja
is currently living in France ... and may want to play under the French
or U.S. flag,” Pahlavanzadeh told the news agency ISNA.Firouzja
wanted to take part in an upcoming world championship in Russia even
though Iran had decided not to attend, Pahlavanzadeh said, without
referring to Israel.Firouzja could not be reached for comment.
In April, Iranian media reported that Firouzja had refused to play against an Israeli player in a tournament in Germany.Iranian
political and sports officials have openly called on the country’s
athletes not to compete against Israelis as a sign of opposition to
Iran’s arch-enemy and solidarity with the Palestinians.Iranian Supreme Leader Ayatollah Ali Khamenei has repeatedly praised athletes who have refused to face opponents from Israel.Since its Islamic Revolution in 1979 Iran has refused to recognize Israel.
In a paper published in the journal Science late last year, Google parent company Alphabet’s DeepMind detailed AlphaZero,
an AI system that could teach itself how to master the game of chess, a
Japanese variant of chess called shogi, and the Chinese board game Go.
In each case, it beat a world champion, demonstrating a knack for
learning two-person games with perfect information — that is to say,
games where any decision is informed of all the events that have
previously occurred.
But AlphaZero had the advantage of knowing the rules of games it was
tasked with playing. In pursuit of a performant machine learning model
capable of teaching itself the rules, a team at DeepMind devised MuZero, which combines a tree-based search (where a tree
is a data structure used for locating information from within a set)
with a learned model. MuZero predicts the quantities most relevant to
game planning, such that it achieves industry-leading performance on 57
different Atari games and matches the performance of AlphaZero in Go,
chess, and shogi.
The researchers say MuZero paves the way for learning methods in a
host of real-world domains, particularly those lacking a simulator that
communicates rules or environment dynamics.
“Planning algorithms … have achieved remarkable successes in
artificial intelligence … However, these planning algorithms all rely on
knowledge of the environment’s dynamics, such as the rules of the game
or an accurate simulator,” wrote the scientists in a preprint paper
describing their work. “Model-based … learning aims to address this
issue by first learning a model of the environment’s dynamics, and then
planning with respect to the learned model.”
Model-based reinforcement learning
Fundamentally, MuZero receives observations — i.e., images of a Go
board or Atari screen — and transforms them into a hidden state. This
hidden state is updated iteratively by a process that receives the
previous state and a hypothetical next action, and at every step the
model predicts the policy (e.g., the move to play), value function
(e.g., the predicted winner), and immediate reward (e.g., the points
scored by playing a move).
Above: Evaluation of MuZero throughout training in chess, shogi, Go, and Atari. The y-axis shows Elo rating.
Image Credit: DeepMind
Intuitively, MuZero internally invents game rules or dynamics that lead to accurate planning.
As the DeepMind researchers explain, one form of reinforcement
learning — the technique that’s at the heart of MuZero and AlphaZero, in
which rewards drive an AI agent toward goals — involves models. This
form models a given environment as an intermediate step, using a state
transition model that predicts the next step and a reward model that
anticipates the reward.
Commonly, model-based reinforcement learning focuses on directly
modeling the observation stream at the pixel level, but this level of
granularity is computationally expensive in large-scale environments. In
fact, no prior method has constructed a model that facilitates planning
in visually complex domains such as Atari; the results lag behind
well-tuned model-free methods, even in terms of data efficiency.
Above: Comparison of MuZero against previous agents in Atari.
Image Credit: DeepMind
For MuZero, DeepMind instead pursued an approach focusing on
end-to-end prediction of a value function, where an algorithm is trained
so that the expected sum of rewards matches the expected value with
respect to real-world actions. The system has no semantics of the
environment state but simply outputs policy, value, and reward
predictions, which an algorithm similar to AlphaZero’s search (albeit
generalized to allow for single-agent domains and intermediate rewards)
uses to produce a recommended policy and estimated value. These in turn
are used to inform an action and the final outcomes in played games.
Training and experimentation
The DeepMind team applied MuZero to the classic board games Go,
chess, and shogi as benchmarks for challenging planning problems, and to
all 57 games in the open source Atari Learning Environment as
benchmarks for visually complex reinforcement learning domains. They
trained the system for five hypothetical steps and a million
mini-batches (i.e., small batches of training data) of size 2,048 in
board games and size 1,024 in Atari, which amounted to 800 simulations
per move for each search in Go, chess, and shogi and 50 simulations for
each search in Atari.
With respect to Go, MuZero slightly exceeded the performance of
AlphaZero despite using less overall computation, which the researchers
say is evidence it might have gained a deeper understanding of its
position. As for Atari, MuZero achieved a new state of the art for both
mean and median normalized score across the 57 games, outperforming the
previous state-of-the-art method (R2D2) in 42 out of 57 games and
outperforming the previous best model-based approach in all games.
Above: Evaluations of MuZero on Go (A), all 57 Atari Games (B), and Ms. Pac-Man (C-D).
Image Credit: DeepMind
The researchers next evaluated a version of MuZero — MuZero Reanalyze
— that was optimized for greater sample efficiency, which they applied
to 75 Atari games using 200 million frames of experience per game in
total. They report that it managed a 731% median normalized score
compared to 192%, 231%, and 431% for previous state-of-the-art
model-free approaches IMPALA, Rainbow, and LASER, respectively, while
requiring substantially less training time (12 hours versus Rainbow’s 10
days).
Lastly, in an attempt to better understand the role the model played
in MuZero, the team focused on Go and Ms. Pac-Man. They compared search
in AlphaZero using a perfect model to the performance of search in
MuZero using a learned model, and they found that MuZero matched the
performance of the perfect model even when undertaking larger searches
than those for which it was trained. In fact, with only six simulations
per move — fewer than the number of simulations per move than is enough
to cover all eight possible actions in Ms. Pac-Man — MuZero learned an
effective policy and “improved rapidly.”
“Many of the breakthroughs in artificial intelligence have been based
on either high-performance planning,” wrote the researchers. “In this
paper we have introduced a method that combines the benefits of both
approaches. Our algorithm, MuZero, has both matched the superhuman
performance of high-performance planning algorithms in their favored
domains — logically complex board games such as chess and Go — and
outperformed state-of-the-art model-free [reinforcement learning]
algorithms in their favored domains — visually complex Atari games.”
In 1957, Mikhail Tal played a training game that sparkled of tactics against his trainer, Alexander Koblents.
One of Tal's most famous chess quotes is "You must take your opponent
into a deep, dark forest where 2+2=5 and the path leading out is only
wide enough for one."
POW! BAM! Raymond Keene, aka The Penguin (er, top), finally gets zapped by the Spectator
SPECTATOR editor Fraser Nelson has at last grasped the nettle so many
of his predecessors – including the current prime minister – were too
timorous to touch. He has booted out the mag’s waddling, bow-tied chess
writer Raymond “the Penguin” Keene, so called because of his uncanny
resemblance to a Batman arch-villain.
Keene’s chess column in the
Speccie two weeks ago had a brief postscript revealing that it was his
last, after 42 years – though it omitted to say why he was going. As a
Spectator source puts it, the Penguin’s offence was “keeping on
plagiarising even after he’d been politely asked not to”.
The Eye has been exposing the Penguin’s bubonic plagiarism, in his books
as well as his columns for the Spectator and the Times, for more than
25 years. In June 1993 (Eye 822) we reported that an entire chapter in
his Complete Book of Gambits had been copied almost verbatim, without
acknowledgement, from an article by American international master John
Donaldson. The publisher had to pay Donaldson $3,000 and withdraw the
chapter from the book, but the waddling word-stealer carried on
regardless.
His most frequently pilfered source was Garry Kasparov’s My Great
Predecessors (2003) – a book he looted so shamelessly he didn’t even
bother to change its distinctive punctuation. Here is Kasparov on a 1954
world championship match: “At the most appropriate moment! By driving
the knight from f6 the pawn spreads confusion in the black ranks.” And a
2009 Spectator column by Keene about the same game: “At the most
appropriate moment! By driving the knight from f6 the pawn spreads
confusion in the black ranks.”
Kleptomania
In 2013 Keene published his Little Book of Chess Secrets – which, as Eye
1357 showed, was largely cut-and-pasted from Kasparov’s My Great
Predecessors. A 2013 Spectator column was also lifted from the Kasparov
book (Eye 1343), reprinting its analysis of a 1923 game between Alekhine
and Rubinstein but pretending it was Keene’s own. For good measure, the
Penguin removed quotation marks and attribution in cases where Kasparov
had quoted other greats, such as Bobby Fischer, thus presenting all the
words as if they were his own.
Keene’s editors can’t pretend they were unaware of his kleptomania. In
October 2008 the chess historian Edward Winter sent a letter by
registered post and email to the Spectator’s then editor: “Dear Mr
d’Ancona, May I advise you that over one third of the chess article by
Mr Raymond Keene published on page 64 of The Spectator, 7 June 2008 was
simply copied, word for word, from what I wrote some two years ago…
Thank you very much in advance for informing me of your proposal for
settling this matter.” Matthew d’Ancona never replied.
Grand larcenist
And how did the Penguin react? By, er, plagiarising another article by
Edward Winter for a Spectator column a year later (Eye 1222), this time
about chess coverage in the Guinness Book of Records. When challenged by
online commenters, Keene affected astonishment. “All I can think of
think of is that somewhere Winter’s comments may have been quoted
without authorship or attribution,” he spluttered, “so I regarded them
as being in the public domain.” Which prompted the question: did he
seriously think it OK to pass off someone else’s article as his own so
long as he didn’t know the author’s name?
In 2013 the chess blogger Justin Horton identified no fewer than 137
columns by the Penguin from the previous three years that included
substantial passages stolen from books and articles by other authors.
“How many more daylight robberies can he get away with,” we asked in Eye
1354, “before the editors of the Times and the Spectator call a halt to
his criminal spree?” Six years on the Spectator has at last got the
message, but Times editor John Witherow still seems unconcerned that his
chess writer is a grand larcenist.
Having been caught red-handed so often, Keene does now occasionally
acknowledge sources – though only fleetingly. Thus in the Times on
Monday 7 October, analysing a 1962 game between Petrosian and Korchnoi,
he said his notes were “based on” those in a book by Dutch grandmaster
Jan Timman. But what does he mean by “based on”? Here is Keene’s
assessment of Korchnoi’s fifth move: “Dubious. With reversed colours
this set-up is fine as the king’s bishop has already been fianchettoed,
although the line won’t yield any advantage. Here, however, the missing
tempo makes itself painfully felt…” And here is Timman’s: “Dubious. With
reversed colours this set-up is OK – since the king’s bishop has
already been fianchettoed – although it won’t yield any advantage then.
But the missing tempo makes itself painfully felt…”
For his Times column the next day, he turned to another Petrosian game
from 1962, against Bobby Fischer. No mention of Timman’s book this time,
but anyone familiar with it would have had severe déjà-vu all over
again. Timman, for example, wrote that after 17 moves “White’s position
is anything but healthy and in the text he manages to save his skin in
an endgame that, with accurate play, he will just about be able to
save.” Keene, by contrast, wrote that after 17 moves “White’s position
is anything but healthy. The text is designed to head for an endgame
that, with accurate play, he should just be able to save.”
Can the beleaguered Penguin save his skin at the Times? Tune in next week, same Bat-time, same Bat-channel.
