Deep learning is useful and will continue to be useful even after AGI gets here. But it is useless to AGI research. In fact, it is a serious hindrance because its practitioners wrongly hail it as a step toward AGI. Thus it is syphoning money away from other promising approaches.
Agree w @ylecun that deep learning best seen as as a tool & should not be dismissed; that’s *exactly* what I concluded Jan 2:
“I don’t think that we need to abandon deep learning.
... we need to reconceptualize it: not as a universal solvent, but simply as one tool among many”https://twitter.com/ylecun/status/952587501037916161 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The analogy doesn't quite pinpoint, for those of us keenly interested,the nature of the inadequacy (aside from it's not being universal).
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Very well said! Choose the method that best fits the data/problem. It is unfortunate DL has come to be viewed synonymously with neural nets. Decouple. Neural nets is not the only DL method. We can view CART trees or deep linguistic parsing without statistics, equally as DL.
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.