Pages

Thursday 7 December 2017

Quick learner

This story has attracted a certain amount of attention but I'm surprised it hasn't attracted more.

BBC version

Google says its AlphaGo Zero artificial intelligence program has triumphed at chess against world-leading specialist software within hours of teaching itself the game from scratch.

The firm's DeepMind division says that it played 100 games against Stockfish 8, and won or drew all of them.

The research has yet to be peer reviewed.

But experts already suggest the achievement will strengthen the firm's position in a competitive sector.

"From a scientific point of view, it's the latest in a series of dazzling results that DeepMind has produced," the University of Oxford's Prof Michael Wooldridge told the BBC.

"The general trajectory in DeepMind seems to be to solve a problem and then demonstrate it can really ramp up performance, and that's very impressive."


There will be unreported caveats and it is reasonable to assume that the two systems ran on different hardware, but on the face of it this achievement looks like a remarkable demonstration of the growing power of AI. Remarkable enough to be disturbing even.

Of course chess is a rule-based environment suited to computation but AlphaGo Zero seems to have taught itself how to play the game to an extremely high standard in a matter of hours.

Strewth was my initial reaction. It still seems appropriate. So much so that I almost hope those caveats douse the whole thing in cold water but I don't think they will.

4 comments:

Sen. C.R.O'Blene said...

Trouble is, people don't have the enormous brain-power to make chess decisions.

Mrs O'Blene and I are pretty-well crack Cribbage players, but there's only a certain algorithm for each card placement. Chess is so different. I played in a torno once, with a chess master. Just one boy beat him out of thirty of us.

Computers take all the fun out of chess, but then so did those fanatics back in the sixties and seventies.

Anonymous said...

I once worked with a chap of whom it was said 'X's intelligence is entirely artificial'. X could talk impressively but was widely regarded as a total dork. Something was missing.

So with current AI, lots of A and little if any I. The difficulty is working out whether the A or the I is the more important component. Indeed does A really need I and could we tell the difference.

Demetrius said...

I can't wait for when DirtyMind comes on stream.

A K Haart said...

Scrobs - I'm not surprised that people still play chess, but surely the days of playing it professionally are numbered.

Roger - with computer chess it is pretty well impossible to tell the difference. If anything chess computers can seem more imaginative.

Demetrius - but will we understand all the innuendo?