How Artificial Intelligence Progresses by Leaps and Bounds

The news in October have come up with this incredible feat: a new version of Google Deep Mind has evolved into the best Go player around learning by itself, and without looking at historical game databases. Some useful links: ‘The AI That Has Nothing to Learn From Humans‘ or ‘DeepMind AlphaGo Zero learns on its own without meatbag intervention‘.

And this happened in just 40 days! This feat was quite impossible to dream only months ago. An engineer from Google is quoted saying “This technique is more powerful than previous versions of AlphaGo because it is no longer constrained by the limits of human knowledge. Instead, it is able to learn tabula rasa from the strongest player in the world: AlphaGo itself.”

It looks like we are entering a new era of Artificial Intelligence at least for activities that are constrained by non-fuzzy logic and for which combinatorial calculation is the key.