- make it possible - make it work - make it efficient - make it dependable and invisible - move on to the next layer and never think about it again
Show this thread
In AI, we're at stage 0 (groundwork). In deep learning, I'd say we're still at the "make it efficient" stage, so efforts to make it invisible & fully-automated may be premature in my opinion. We still need the monthly power consumption of a village to "solve" CIFAR10 from scratch
-
-
Efforts to make it "INVISIBLE & FULLY-AUTOMATED" would be horrendously immoral, given the stage we are currently in.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Do you really think that in DL we are at the make it efficient stage? It seems to me that we are still exploring what neural architectures make the most sense for certain domains (and that’s not exactly an issue of efficiency).
-
We're still exploring the space of problems we can solve, but I think that was mostly 2014-2016 (while 2011-2013 was "wow, it works"). DL is now a well-established solution for a certain scope of problems, but still a very inefficient one (data-inefficient, power-inefficient...)
- Show replies
New conversation -
-
-
Machine learning has an advantage over human learning there. You can copy weights between machines. Babys have to experience all the pain from scratch every single time.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Some aspects of the fully-automated stage will have value in the "make it efficient" stage. e.g., Google Research papers showing that automated architecture search discovers architectures with orders of magnitude better efficiency + similar/better performance than human-designed.
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.