Machine Learning: Powerful Virtuoso

When DeepMind’s AlphaZero program played against the reigning champion of chess-playing programs, it sounds a bit like an ancient Roman gladiator going head-to-head with a modern-day marine. AlphaZero made a rather confusing retreat of its queen into the corner:

Yet this peculiar retreat was venomous: No matter how Stockfish replied, it was doomed. It was almost as if AlphaZero was waiting for Stockfish to realize, after billions of brutish calculations, how hopeless its position truly was, so that the beast could relax and expire peacefully, like a vanquished bull before a matador. Grandmasters had never seen anything like it. AlphaZero had the finesse of a virtuoso and the power of a machine. It was humankind’s first glimpse of an awesome new kind of intelligence.

One Giant Step for a Chess-Playing Machine

AlphaZero represents an example of a new type of intelligence that I like to describe as “statistical thinking.” This new intelligence draws data-driven conclusions backed by massive volumes of data. It uses the raw inorganic processing power of machines to drive elegant pattern-matching techniques, some inspired by nature. The result is a new kind synthetic vision, akin perhaps to the telescope and microscope, that gives us new insight into the underlying behavior of physical reality. Or, as former chess champion Gary Kasparov describes it:

By discovering the principles of chess on its own, AlphaZero developed a style of play that “reflects the truth” about the game rather than “the priorities and prejudices of programmers.”

One Giant Step for a Chess-Playing Machine

Image credit: Ahn Young-joon/Associated Press

Scroll to Top