4 ms·
Still other way of looking at this - could the whole humanity playing in one team win against AlphaGo? If no - computers still win because they scale better (p
by ajuc 11y ago
Still other way of looking at this - could the whole humanity playing in one team win against AlphaGo?
If no - computers still win because they scale better (play better as a team).
- sarreph 11y agoLet's set the entire world population up as a huge deep neural network (each person acting as a node) and put that theory to the test! ;)
- smitherfield 11y agoAlphaGo is the product of the combined human effort to play Go as well as possible.
- pmontra 11y agoIMHO the whole humanity won't contribute much to the challenge. You take the best 5 pros and that's pretty much the best team you can pit against the AI. Ten pros would probably need too much time to discuss and coordinate. A striking analogy with software dev teams :-)
- ajuc 11y agoThat's the point - humans don't scale, and 1 person vs energetic equivalent computer is arbitrary (and human-prefering) level of comparison.
- visarga 11y agoOr if they played 5 second Go, that would reduce the time the computer can do Monte Carlo tree search, which is the brute force aspect of AlphaGo.
- DennisP 11y agoSomething like that was actually tried in 1999, when Kasparov played against the world: http://michaelnielsen.org/blog/kasparov-versus-the-world/ http://michaelnielsen.org/blog/kasparov-versus-the-world/ Given how well collective intelligence worked in that game, it'd be pretty interesting to try it against AlphaGo.