Based on a note by Florian Forestier
The recent clash between the AlphaGo computer programme developed by DeepMind, Google’s artificial intelligence arm, and the South Korean Go champion and world number three Lee Sedol has, understandably, been described as a stage in a single drama leading inexorably to the coming of a ‘strong’ Artificial Intelligence (AI), which is highly likely to equal human beings and then surpass them in all their activities.
Even more than the power of AlphaGo, what has interested commentators is its mode of learning. It is neither through sheer processing power nor even a statistical...