Mankind's last outpost
Location: South Korea
Date: 8-15 March 2016.
A machine makes a move that no human ever would. “Divine” is the word. In a huge breakthrough,
the world recognizes today the pinnacle—and somewhat alarming—height reached today by
Artificial Intelligence.
In China, there is a game that has been played for over 2500 years. To play it, two players
alternately place black and white pawns on the empty intersections of a 19 × 19 checkerboard,
called “goban.” The aim of this game is to control a zone of the goban greater than the one
controlled by the opponent. We are talking about "Go", perhaps the most complex game on earth,
judging by the magnitude of the playing field, and therefore of possible combinations. Playing Go is
also an excellent way to develop both the creative and the logical abilities because both sides of
your brain need to be activated during the game.
In this week of May, it’s going on an epic battle between Lee Sedol and “AlphaGo,” an AI system
designed by a squad of researchers at DeepMind, a London AI lab now owned by Google.
Lee Sedol is a professional Go player, and he is to Go what Mozart is to music. Indeed, at the young
age of 12, he was promoted to professional Dan rank, winning 18 world championships since then.
He is a living legend in his native South Korea, where he is known for his unconventional and
creative play.
The prize for this contest is one million dollars, at best five matches. Actually, the stakes of this
game far exceed a fistful of dollars. Go is man’s latest outpost, now that computers are dominating
any other table game requiring both logical and strategic skills.
To many, the match is no more than a good workout for the Korean champion, who promises to
knock-out his silicon opponent without breaking a sweat. Even a victory by a narrow margin would
be a shame for the Korean hero, based on the complexity of the game, which can only be tamed by
a thinking being.
In reality, the match was actually a walk in the park, but for his opponent that claimed victory
winning four games and losing only one. It marked the first time a machine had beaten the very best
player at this ancient and enormously complex game; a feat that even hopeless optimists didn't
expect before 2025.
Behind the curtain, at the heart of AlphaGo, there is today's best-of-breed of Artificial Intelligence
(AI). AlphaGo relies on deep neural networks, which are a network of software and hardware
components that can mimic the human brain capabilities.
Much like the human brain, neural networks can gather information and react to it. They can build
up an understanding by analyzing massive amounts of digital data. If you feed enough photos of a
cat to the neural network, it can learn to recognize a cat. Accordingly, if you show how to play Go
to a neural network, it can learn the game of Go. It can be so smart that it can even start playing a
match against itself, and learn from it. Artificial Intelligence is not, however, something available only in the realm of supercomputers.
We lesser mortals can also enjoy a taste of it every day. Whoever has got (who hasn’t?) a modern
smart-phone should be aware of digital personal assistants which can help you find useful
information just using your voice; you may say “Where’s the nearest Italian restaurant?”, “What’s
on my schedule tomorrow?”. You might be surprised to know that even Santa Claus’ secret is in
safe hands should your kid ask him. Mine did it.
Sometimes the road to Artificial Intelligence is paved with false starts: for example, a while ago
Google Map included a feature that aimed at encouraging people to walk instead of driving, by
showing how many calories one could burn, and what one could eat by that amount of calories.
I personally found that feature sort of amusing! Nevertheless, it was banned for being “not ethical”
as it could play havoc with people suffering from eating disorders. Too bad nobody will tell me
anymore how many cupcakes I’m allowed to eat when I cover my distance up to my car.
AI, however, is not just for fun: while many have blind faith in the underlying value of these
technologies, it’s hard to escape the feeling that it’s scary at the same time. The primary concern is
that super-intelligent machines will bring about the end of human work and maybe even break free
from our control.
Does this sound familiar to Hollywood fans?
Date: 8-15 March 2016.
A machine makes a move that no human ever would. “Divine” is the word. In a huge breakthrough,
the world recognizes today the pinnacle—and somewhat alarming—height reached today by
Artificial Intelligence.
In China, there is a game that has been played for over 2500 years. To play it, two players
alternately place black and white pawns on the empty intersections of a 19 × 19 checkerboard,
called “goban.” The aim of this game is to control a zone of the goban greater than the one
controlled by the opponent. We are talking about "Go", perhaps the most complex game on earth,
judging by the magnitude of the playing field, and therefore of possible combinations. Playing Go is
also an excellent way to develop both the creative and the logical abilities because both sides of
your brain need to be activated during the game.
In this week of May, it’s going on an epic battle between Lee Sedol and “AlphaGo,” an AI system
designed by a squad of researchers at DeepMind, a London AI lab now owned by Google.
Lee Sedol is a professional Go player, and he is to Go what Mozart is to music. Indeed, at the young
age of 12, he was promoted to professional Dan rank, winning 18 world championships since then.
He is a living legend in his native South Korea, where he is known for his unconventional and
creative play.
The prize for this contest is one million dollars, at best five matches. Actually, the stakes of this
game far exceed a fistful of dollars. Go is man’s latest outpost, now that computers are dominating
any other table game requiring both logical and strategic skills.
To many, the match is no more than a good workout for the Korean champion, who promises to
knock-out his silicon opponent without breaking a sweat. Even a victory by a narrow margin would
be a shame for the Korean hero, based on the complexity of the game, which can only be tamed by
a thinking being.
In reality, the match was actually a walk in the park, but for his opponent that claimed victory
winning four games and losing only one. It marked the first time a machine had beaten the very best
player at this ancient and enormously complex game; a feat that even hopeless optimists didn't
expect before 2025.
Behind the curtain, at the heart of AlphaGo, there is today's best-of-breed of Artificial Intelligence
(AI). AlphaGo relies on deep neural networks, which are a network of software and hardware
components that can mimic the human brain capabilities.
Much like the human brain, neural networks can gather information and react to it. They can build
up an understanding by analyzing massive amounts of digital data. If you feed enough photos of a
cat to the neural network, it can learn to recognize a cat. Accordingly, if you show how to play Go
to a neural network, it can learn the game of Go. It can be so smart that it can even start playing a
match against itself, and learn from it. Artificial Intelligence is not, however, something available only in the realm of supercomputers.
We lesser mortals can also enjoy a taste of it every day. Whoever has got (who hasn’t?) a modern
smart-phone should be aware of digital personal assistants which can help you find useful
information just using your voice; you may say “Where’s the nearest Italian restaurant?”, “What’s
on my schedule tomorrow?”. You might be surprised to know that even Santa Claus’ secret is in
safe hands should your kid ask him. Mine did it.
Sometimes the road to Artificial Intelligence is paved with false starts: for example, a while ago
Google Map included a feature that aimed at encouraging people to walk instead of driving, by
showing how many calories one could burn, and what one could eat by that amount of calories.
I personally found that feature sort of amusing! Nevertheless, it was banned for being “not ethical”
as it could play havoc with people suffering from eating disorders. Too bad nobody will tell me
anymore how many cupcakes I’m allowed to eat when I cover my distance up to my car.
AI, however, is not just for fun: while many have blind faith in the underlying value of these
technologies, it’s hard to escape the feeling that it’s scary at the same time. The primary concern is
that super-intelligent machines will bring about the end of human work and maybe even break free
from our control.
Does this sound familiar to Hollywood fans?
