Google’s AlphaGo wins first match against Chinese go top player

Google’s artificial intelligence AlphaGo has won its first of three matches against the Chinese Ke Jie. Nineteen-year-old Jie is currently the highest ranked go player, but he was expected to lose.

AlphaGo won with only half a point difference, the smallest margin there is, thus Demis Hassabis, the CEO of DeepMind, which developed AlphaGo. Hassabis claims that Jie used ‘ideas’ that AlphaGo used during previous matches. According to The Verge, artificial intelligence has been winning from the start. The second match will follow on Wednesday during the Future of Go Summit in Wuzhen, China, and the virtual and human players will meet for the final on Saturday.

The competition takes place more than a year after AlphaGo’s victory over Korean go world champion Lee Sedol. Since artificial intelligence has learned a lot since then, the win over Ke Jie comes as no surprise, Wired writes. In between, AlphaGo played online against top go players, including Jie. In doing so, he won his first fifty matches and drew in the 51st.

AlphaGo is getting better and better through the use of techniques such as supervised learning, in which predictions are repeatedly made based on data sets, and reinforcement learning, in which a neural network continuously improves its rules.