When you look back and squint, you can see just how bizarre it is that data-hungry ML algorithms have spawned an entire industry creating labeled data to feed them. We’ve somehow put the cart before the horse.
Show this thread
-
The ridiculousness of the situation is compounded in the cases where those labels require subject matter expertise. That means that either we are training a whole new generation of SMEs to create data, or more often, having non-experts make their best guess.
Show this thread
ML is cognitive automation, so it needs to be provided with an original template to automate. Can't photocopy a document you don't have.
-
-
Absolutely. It’s been wild to watch the paradigm evolve from “generalize from observations” to “create a curriculum to educate the algorithm.”
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
ML wants to be "cognitive automation" as you call it. Today it's anything but - we keep finding out it's just "automation we don't understand"https://twitter.com/ShashankaMadhu/status/1389981121555976205?s=20 …
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.