There's a big difference between learning to solve problems on your own and learning to look up existing solutions. If you want to unlock your potential, learn the former. If you want to unlock AI's potential, teach it the former. The goal is to memorize as little as possible
-
-
The same is often true with research: try to reinvent first, look up the literature afterwards. If you want to be able to produce original thoughts, you need to create space for them to develop, instead of always filling up empty pockets of concept space with ready-made solutions
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I think there's value to getting good at applications first, then back-filling in the theory. When you have experience with training NN's already, you'll have a much better mental scaffold for understanding the math behind backprop, activation functions, etc.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is the only sustainable method. Unless you don't want to challenge yourself going forward, pattern-matching eventually falls off a cliff. Good reminder.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
My philosophy for the Rosalind problem set, which makes for slow progress, but the knowledge you get from bootstrapping your own solution can’t be gotten another way.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
There’s also the middle ground of analyzing worked examples. That can give you a deeper understanding than rotely memorizing the solutions while also reducing the time and mental noise of hitting your own dead ends.https://en.wikipedia.org/wiki/Worked-example_effect …
-
That makes sense, though one could argue you would not be as prepared to face dead ends in your bleeding edge research endeavors

- Show replies
New conversation -
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.