Forget about Chess
Ever heard about the board game Go? With a 3,000-year history, this board game is far more complex than chess. That said, some say there are more moves in chess than the number of quarks in the universe. Yet, a Computer Program named Deep Thought beat an International Grand Master for the first time in November 1988 in Long Beach, California. That was 29 years ago. Go on the other hand has been a harder nut to crack, with a supposed number of moves of 10^800 (a 10 with 800 zeroes after it) for every atom in the known universe. Since that fateful November day in 1988, machines had yet to beat a human at their own game.
DeepMind is a neural network developed by a company by the same name, which happens to be a subsidiary of Google (Alphabet). It is designed to learn from experience, much like a human. AlphaGo is an AI module, utilizing the neural network to play Go. Through feeding it with data from countless Go outcomes, the AI was able to beat the world champion Ke Jie 23-27 May this year in an official three-game match. As a reward for beating the grand master, AlphaGo was awarded with the same professional 9-dan rank as Ke by the Chinese Weiqi Association.
Now the researchers are aiming for more than just games.
The team behind AlphaGo said they would retire from the project; instead, focusing on different AI projects. They didn’t name any specifics, however.
The future of AI
Imagine if a computer could give you an accurate diagnosis based on your symptoms and lab data? The precision would be far above that of a human doctor and the computer could even be able to calculate your drug dosage based on your genetics, body weight etc. A neural network has the potential to do just that, and more. Games are just the first step of ultimately bringing true artificial intelligence to our civilization. Influential critics like Stephen Hawking and Elon Musk say AI will be the end of humanity, given the AI start to truly think for itself.
What do you think? Leave a comment and let us know below!