Machine Learning Conquers the real-time strategy game StarCraft II
ARTIFICIAL INTELLIGENCE
CONQUERS STARCRAFT II IN 'UNIMAGINABLY UNUSUAL' AI BREAKTHROUGH
DeepMind was behind other
major breakthroughs with games like Go
Anthony Cuthbertson
Wednesday 30 October 2019 19:13
A major artificial
intelligence milestone has been passed after an AI algorithm was able to defeat
some of the world's best players at the real-time strategy game StarCraft II.
Researchers at leading AI
firm DeepMind developed a programme called AlphaStar capable of reaching the
top eSport league for the popular video game, ranking among the top 0.2 per
cent of all human players.
A paper detailing the
achievement, published in the scientific journal Nature, reveals how a
technique called reinforcement learning allowed the algorithm to essentially
teach itself effective strategies and counter-strategies.
"The history of
progress in artificial intelligence has been marked by milestone achievements
in games. Ever since computers cracked Go, chess and poker, StarCraft has
emerged by consensus as the next grand challenge," said David Silver, a
principal research scientist at DeepMind.
"The game's complexity
is much greater than chess, because players control hundreds of units; more
complex than Go, because there are 1026 possible choices for every move; and
players have less information about their opponents than in poker."
DeepMind, which was acquired
by Google in 2014, was behind the first ever AI algorithm capable of beating a
human champion at the ancient Chinese board game Go.
In 2016, the firm's AlphaGo
program defeated grandmaster Lee Sedol, who is recognised as the best player in
the world. The pioneering match finished 4-1, however improvements to the algorithm
have seen new generations of AlphaGo defeat the original version.
A modified version of
AlphaZero has since gone on to master other two player games like chess and
shogi to a superhuman level.
"StarCraft has been a
grand challenge for AI researchers for over 15 years," said DeepMind
co-founder Demis Hassabis. "These impressive results mark an important
step forward in our mission to create intelligent systems that will accelerate
scientific discovery."
Professional StarCraft
players described AlphaStar's technique as unusual but on a similar level to
the very best human players of the game.
"AlphaStar is an
intriguiing and unorthodox player - one with the reflexes and speed of the best
pros but strategies and a styled that are entirely on its own," said Diego
'Kelazhur' Schwimmer, a professional StarCraft II player for eSports team Panda
Global.
"The way AlphaStar was
trained, with agents competing against each other in a league, has resulted in
gameplay that's unimaginably unusual... Though some of AlphaStar's strategies
may at first seem strange, I can't help but wonder if combining all the
different play styles it demonstrated could actually be the best way to play
the game."
Comments
Post a Comment