1/ Important lesson from AlphaGo: as AI gets better than humans, it gets very hard to evaluate.
-
2/ Experts thought Sedol and AlphaGo were close at the beginning of the games, but maybe they never were and we just didn't understand.
@sama The AlphaGo team almost certainly knows - they'll have data on the estimated probability of winning through the match.
-
-
@sama Relatedly, the Deep Blue paper contains many diagnostic tidbits about what actually mattered, eg (one of many)pic.twitter.com/y7bzZ7hIkt
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@michael_nielsen@sama is that probability based on what AlphaGo thinks or how we measure what we think AlphaGo thinks?Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@michael_nielsen@sama That estimate is incorrect. The team mentioned that it was essentially impossible to know by how much.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.