An article by: David Tan, Chief Technology Officer, CHIPS Technology Group
Go is a very simple board game – on the surface at least. Many of us played it as kids. The game consists of a large grid (like chess or checkers), and a series of pieces with black on one side and white on the other. The goal is simply to surround your opponent, thus turning all their pieces to your color, and eventually control the entire board. It’s an ancient Chinese board game that dates back almost 3000 years, and yet today it stands as the symbol of the future of Artificial Intelligence.
Despite the
apparently simple rules, the strategy involved in the game is quite
complex. The main reason for that is the
number of possible different games.
Chess, a traditionally very complex game of strategy, has 10 to the 120th
power possible game combinations. Go has
10 to the 761st power. The
level of complexity has made it prohibitively difficult for computers to win at
over humans. IBM built a super computer
that could beat a chess master almost 20 years ago, but we still couldn’t build
a computer to beat a master Go player.
Until today.
Last week.
Google pitted AlphaGo, a program developed on their DeepMind artificial
intelligence platform against legendary Go player Lee Se-dol. AlphaGo used complex machine learning and
neural networks to evaluate and execute strategy at a world-class level, and
for the first time ever, a computer beat a master champion in a game without
handicaps. In fact, the computer won 4
out of 5 games, leaving Se-dol stunned at the outcome.
The victory
signifies a major milestone in artificial intelligence research. Most experts would have estimated that an
intelligent machine capable of achieving this type of victory was at least 5
years away, if not 10. Even as the
matches started, most expected the human to come out on top. Even DeepMind’s creators were stunned at the
level of their success. In some ways,
this marks the end of a 20-year journey started when IBM’s Deep Blue beat Gary
Kasparov at chess. Mastering board games
is no longer a measure of the power of computers. It’s time to move on to bigger challenges.
It’s
important to remember that AlphaGo was programmed for the express purpose of
mastering Go, but the underlying technology is much more multipurpose. Google didn’t invest in the company and build
up the technology to build a better game playing machine. They have big plans for the power of AI, and
how it can be applied to big real-world problems.
AI is
already being used in some incredible ways that you might not even
realize. IBM’s Watson for example can
“read” text and answer questions about it.
It understands subtlety and nuance.
Watson is being used extensively in healthcare around the world to aide
doctors in the understanding and treatment of all kinds of conditions. Watson is even learning to “see” so it can
spot anomalies on x-rays or MRIs for example before a doctor could.
I bet you
don’t even realize some of the places you come across AI in your everyday
life. Google for example uses a
deep-learning neural network called RankBrain as one of their key signals for
search result rankings, which it credits for tremendous improvements in search
over the last 2 years. Many of the news
articles you read aren’t being written by humans anymore. The Associated Press for example uses an AI
platform to automatically write all its earning reports – and I bet you can’t
tell the difference. Ever chat online
with an automated shopping assistant?
Chances are that was a machine powered by AI and natural language
processing. The bottom line is, this
technology is here and already changing industry.
The final
frontier for AI is of course robotics, and there are many companies at work in
this space. While it’s still somewhat
off in the future, much progress is being made.
Companies like Google and IBM continue to pour billions of R&D
dollars into this work. Just this week,
IBM announced a pilot program with Hilton Hotels for a robot concierge that can
sit in the lobby, answer your questions, make reservations and give you
directions. It’s still the early stages
of what will ultimately be available, but the power and possibilities is
incredible.
Artificial
Intelligence will transform every industry and the very way we live. Some fear for the rise of the machines, and
of course we need to keep control of this technology to be sure, but there is
no reason for paranoia. Incredible
advances in science, technology and healthcare will be nothing
but good for the world. The machines are
getting smarter, and they’re not playing games anymore. And that’s a good thing.