Go. Image: Saran Poroong/Shutterstock
Google’s artificial intelligence arm DeepMind took the historic game of Go from humanity’s grasp last year. It plans on underlining its dominance once again in 2017.
The power and performance of AI reached new levels in 2016 when, before many people knew it was even possible, a computer program won one of the most complex board games in existence.
Over the course of a few weeks and five games, Lee Sedol – champion Go player, rated as one of the best ever in a game stretching back 3,000 years – was taken apart by AlphaGo.
Google’s DeepMind AI operation created AlphaGo with the sole aim to master the field and, in beating Sedol, it did so.
As if to undermine its status as the best on the planet, AlphaGo is back this year, and the world’s current number one Go player, Ke Jie, is the latest potential lamb to the slaughter.
However, Demis Hassabis, co-founder and CEO of DeepMind, maintains that 2017’s battle won’t be another nail in humanity’s Go-playing coffin. Rather, it will help humans to perform even better.
“Instead of diminishing the game, as some feared, AI has actually made human players stronger and more creative,” he said.
“It’s humbling to see how pros and amateurs alike, who have pored over every detail of AlphaGo’s innovative game play, have actually learned new knowledge and strategies about perhaps the most studied and contemplated game in history.”
The Future of Go summit in China, which may have an air of dread around it, will see AlphaGo go up against Jie, as well as a series of other top players across multiple formats of the game, to show just how far along computers have come.
One format will see five human players take on AlphaGo together. Another will see humans partner with AlphaGo to beat other humans, also partnered with the computer program.
The ancient Chinese game has probably never seen such variety in how it is played.
Of course, Go is not the first game to be mastered by AI. Deep Blue finally beat Garry Kasparov, widely thought of as one of the world’s best chess players, in 1997.
What added to that drama was that Deep Blue was returning after losing to the same man in 1996. Sadly, a sporting trilogy was never achieved in 1998.
Elsewhere, Libratus, an improved AI program based on a previous one called Claudico from 2015, beat four professional poker players at Heads-Up No-Limit Texas Hold’em.
The ‘Brains v Artificial Intelligence’ repeat saw four human players face off against Libratus: Dong Kim, Jimmy Chou, Daniel McAulay and Jason Les. The latter finished 40th in last year’s World Series of Poker main event.
They played separate games with Libratus, each logging 120,000 hands against the machine and ultimately losing by a cumulative $1.7m in chips. After four days, the margin was $50,000 but, from there, the algorithm proved its worth and the pros were crushed over three weeks.
