Thanks Gary. We don't go deep into the broader picture, but in general I think this is a good non-algorithmic way to incorporate human intelligence in AI. Human-defined subset testing can avoid many bone-headed AI failures, at least in constrained and understood environments.
important, thoughtful essay on medical AI techniques and how to apply them more safely, by @DrLukeOR w @HazyResearch and others.https://twitter.com/DrLukeOR/status/1183708204749869056 …
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Simple scores are terrible for denoting performance in real world. They neither capture the degree or the gravity of the error.
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.