In other words, the argument "deep learning is overhyped; there are things it will remain fundamentally unable to do because it's just a GFR" has a specific meaning, it's not handwaving that would apply to any learning system
Show this thread
Deep learning is a giant fancy regression (GFR). But lots of things aren't. Most software running on your computer isn't a GFR. Large chunks of the field of AI isn't about GFR. And your brain isn't a GFR (though it may implement something like it as a small part of what it does).https://twitter.com/Noahpinion/status/990820050968788993 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
So, in summary: You better have a Good Friggin Reason for machine learning?
-
It's highly unlikely that this is an appropriate summary.
- Show replies
New conversation -
-
-
I’d already been a fan of unification of models in the
#AI space but after starting my first read thru#TheMasterAlgorithm I’m even more intrigued. DL is not the one great solution but it is certainly part of the solution.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
If by regression one refers to regressing outcomes on predictors, well, maybe. In reality of course deep learning has nothing to do with OLS.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I still believe that it is a brute Force memorization techniques. The semantics at play for memorization and learning does not change this.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Is that the case for deep reinforcement learning as well?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What about giant fancy parameter search?
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.