Is robustness aggainst adversarial input included in FDA approval procedures? Do specialized FDA approval procedures for DL-based diagnostics exist at all?
I had earlier tweeted about the faulty data being used for #AI based skin cancer detection. Here's more details from domain expert. #DeepLearning is bound to fail in Healthcare. No one wants blackboxes with spurious correlationshttps://twitter.com/Farzad_MD/status/1170332477514276864 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Totally! You think that's bad, think of this: folks started thinking that
#AutoML is a replacement for rigorous training! All they need is good problem! -
New conversation -
-
-
>
#DeepLearning is bound to fail in Healthcare. So interesting. Anima, do you think DL could be improved somehow to make it suitable here, or not? What is the right approach.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.