It's been what, 5 years now? I think its high time we start asking questions.
Has deep learning solved atleast one problem? One problem where it has really cracked open the solution.
In most cases I see regular learning approaches work.
cc @filippie509 @GaryMarcus
-
Computer Go. I realize it's not a business, but it's been solved. Similarly, computer chess won't go back to the days of tree search based methods any time soon. Deep learning is here to stay in these areas. Of course, these are not businesses.
-
-
Yes, tree search is essential to alpha*. As is DL. Only an intelligent combination of both MCTS and DL provides a state of the art solution. Both are equally important here.
-
yes, that’s why i have been arguing for hybrid models for over 25 years, and have made that specific point wrt to AlphaGo since it was first reported.
- 1 more reply
New conversation -
-
-
and similarly a key part of the Rubik’s cube solver that was just published.
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.