‘Go’, an ancient Chinese board game, was considered by DeepMind’s CEO to be ‘the only game left above chess’, which computers conquered in 1997 when IBM’s Deep Blue defeated the world champion. The game seems simple, but there are trillions of possible moves, which makes it difficult to follow a particular strategy. Mastering the game means using intuition to react to the twists and turns that appear.
Many had believed that it would be at least another decade before artificial intelligence developed enough to conquer ‘Go’, but AlphaGo has proven them wrong. Its mastery of the game is so significant to AI development because of the reliance the top players have on intuition as they play the game. DeepMind built ‘reinforcement learning’ into the programme, which measnt hat the machine plays against itself and adjusted its own neural networks based on trial and error. The program is capable of narrowing down the search space for the next move from the near-infinite and of anticipating the long-results of each move.
Computer Science applicants should look more into AI and the programming used on Alpha Go that has led to its victory, while Mathematics students should investigate the mathematical side of the board game, and how a computer can better calculate the best possible move than the human mind.
In a shock turn of events, Lee Sedol finally beat AlphaGo in their fourth match against each other. Those watching, including the CEO of DeepMind, believe that the computer did make some mistakes, prompting the question of whether AI will ever truly outsmart the human mind, or whether this loss will only help the program to grow stronger. Students applying for Anthropology or HSPS should consider the wider implications of the development of AI, and how this may affect the society we live in.