By: Charlie P.
Have you ever dreamed of an era where robots are doing everything for you from cooking to teaching? Well, we may now be closer than ever to achieving this, thanks to Google’s DeepMind Technologies’ development of AlphaGo. AlphaGo is a computer program which was specifically developed to play the ancient Chinese board game “Go” using artificial intelligence, which has beaten the Chinese world champion Ke Jie in May 2017.
What is Go, and how do we play it?
Go, as mentioned before, is a board game which was first invented more than 2500 years ago in China. It involves two players and the playing pieces are called “stones”. Black stones are assigned to one player and white stones are assigned to the other. The game is played on a board where there are 19 horizontal and vertical lines, and the players make moves on the intersections of the lines by taking turns. The player with black stones start first, and the main goal of the game is to conquer as much “territory” as possible, meaning that the player has to surround or contain the maximum area possible using their stones. Despite its simple objectives, the game is considered to be one of the most complex games on Earth because of the fact that there are about 10^170 possible moves, which is a greater amount than the total quantity of observable atoms in the entire universe.
Why is this AlphaGo significant?
AlphaGo is certainly significant because it is one of the first machines that truly uses artificial intelligence, and in this case, the program is learning how to play by itself. Unlike the DeepBlue which beat the chess world champion in 1997, and plays chess by calculating the best possible move, AlphaGo functions similarly to a person’s brain. According to Demis Hassabis, the CEO and Co-founder of DeepMind, AlphaGo first analyzes millions of Go games played by professionals, and then plays itself thousands of times to discover new strategies in order to beat other professionals. In 2016, its system was proven successful by beating the legendary Korean Go player Lee SeDol 4-1. However, the fact that it did not win 4-0 showed that the program was not perfect, and therefore the decision was made to improve it. In May 2017, the program was proven perfect when it beat the number 1 Go player in the world, Ke Jie, 3-0.
20 years passed from the development of a chess program to a Go program, and it may feel as if we still have a long way to go. However, great advancements have already been achieved, such as the development of virtual reality, and the 3D printer. Moreover, it is rumored that
Google’s Waymo is testing self-driving trucks. Through the development of AlphaGo, we might have just witnessed the dawn of the age of artificial intelligence, in which it exceeds human performance and starts to replace it.