Probably every aspect and every detail about AlphaGo and Lee Sedol have been covered since the incredible matches. Yet, many questions are left to be answered inside and outside of the go board. Here, we have the pleasure of hearing the insight of an active professional player, Zhao BaoLong 2p after he studied these games:
- AlphaGo is superior to human in the whole board thinking. That is to say, when judgement is involved, the computer preforms better.
- Seen from these 5 games, human can compete with the computer in calculation.
- The computer does not choose the “best” moves, but the “safest” moves. When it is leading, it is unlikely for human to turn the game around.
- The computer tries moves that are unthinkable for human when it is behind.
- It seems there is still flaw in AlphaGo, the question is: can we find it?
As a professional player, this new development did not only largely promoted Go in the world but also brought a new era of Go. With the inspiration of computer, many of the concepts with which we are familiar will probably be changed completely. This of course enables us to be free of our limitations by these concepts.
Zhao BaoLong 2p is very kind to share his study with us, more detailed comments of these five games can be found in the sgf files below.