Artificial intelligence beats eight world champions in bridge | Artificial Intelligence (AI)

An artificial intelligence has beaten eight world champions at bridge, a game where human supremacy has so far resisted the march of machines.

This victory represents a new step for the AI ​​because in bridge, players work with incomplete information and must react to the behavior of several other players – a scenario much closer to human decision-making.

By contrast, in chess and Go – in which AIs have already beaten human champions – a player has only one opponent at a time and both are in possession of all the information.

“What we have seen represents a fundamentally important advance in the state of artificial intelligence systems,” said Stephen Muggleton, professor of machine learning at Imperial College London.

French startup NukkaAI announced the news of its AI victory on Friday, following a two-day tournament in Paris.

The NukkAI challenge required human champions to play 800 consecutive deals divided into 80 sets of 10. It did not involve the initial bidding component of the game in which players arrive at a contract which they must then fulfill by playing. their cards.

Each champion played their own cards and those of their “fictitious” partner against a pair of opponents. These opponents were the best bot champions in the world to date – bots who have won numerous bot competitions but are universally acclaimed to be nowhere near as good as expert human players.

The AI ​​- called NooK – played the same role as the human champion, with the same cards and the same opponents. The score was the difference between those of the human and the AI, averaged over each series. NooK won 67, or 83%, of 80 sets.

Jean-Baptiste Fantun, co-founder of NukkaAI, said he was confident the machine – which the company has been developing for five years – would triumph in thousands of transactions, but with just 800 it was touchscreen.

Announcing the results, mathematician Cédric Villani, winner of the Fields Medal in 2010, called NukkaAI a “superb French success story”.

Artificial intelligence researcher Véronique Ventos, the other co-founder of NukkaAI, calls NooK a “next-gen AI” because she explains her decisions as she goes. “At bridge, you can’t play if you don’t explain,” she says.

The game is based on communication between partners.

Explainability is a hot topic in AI. “Most of what the general public has heard in recent years about machine learning is based on black box systems such as AlphaGo, which are incapable of explaining to human beings how decisions are made,” said Muggleton.

Instead, NooK represents a “white box” or “neurosymbolic” approach. Rather than learning by playing billions of rounds of a game, he first learns the rules of the game and then improves his game through practice. It is a hybrid of rule-based systems and deep learning. “The NooK approach is learning in a much closer way to human beings,” Muggleton said.

“The pendulum swings toward these kinds of methods,” says Michael Littman, professor of computer science at Brown University in Rhode Island. “Not being able to tell people what’s going on just doesn’t work in our societies.”

Even if a person or an AI can’t explain in words what they’re doing, Littman says, their behavior needs to be “readable” to others, making rules they understand.

This will be essential in areas such as healthcare and engineering. Self-driving cars negotiating a crossroads will need to be able to read each other’s behavior, for example.

Littman was disappointed the challenge didn’t include auctions, where most of the most interesting communications — and deceptions — occur in bridge.

But Nevena Senior, a multiple world bridge champion for England and one of NooK’s challengers, said the contracts the humans and NooK had been given were variable enough for the card game to become as big as auctions.

She said the creators of NooK did a “magnificent” job. She discovered that he read his opponents better than humans and was better at exploiting their mistakes.

“It’s something humans do after enough experience and I was pleasantly surprised that a robot mimics typical human skills,” she said.

Other AI Triumphs

  • 1996: IBM’s Deep Blue chess machine wins a game against world chess champion Garry Kasparov but loses the match 2-4. A year later, Kasparov lost the rematch.

  • 2007: Checkers is solved by researchers at the University of Alberta in Canada. After sifting through 500 billion positions, they build an unbeatable checkers game computer program.

  • 2011: IBM’s Watson computer beats the game show Jeopardy! champions Brad Rutter and Ken Jennings, winning the $1 million top prize.

  • 2016: Google DeepMind’s AlphaGo defeats Korean Go champion Lee Sedol 4-1. The Korea Baduk Association awards AlphaGo the highest rank of Grandmaster of Go, an honorary 9 dan.

  • 2022: NooK, NukkaAI’s bridge computer, defeats eight world bridge champions in Paris.

Leave a Comment