Also E.g. an ImageNet model will be superhuman with dog types but classify people as seatbelts, and that’s okay - working as intended.
@karpathy No child would ever confuse person with seatbelt, even w small data. For open-ended vision we need more than current tech offers.
-
-
@GaryMarcus@karpathy Agree Gary -
@marek_rosa@GaryMarcus goes to my point. misunderstanding of ImageNet's labels. There's no person class. Seatbelt class has many ppl. - 3 more replies
New conversation -
-
-
@GaryMarcus@karpathy True, but can we get there from here? -
@AlanMackworth@karpathy No cognition without representation? Causation + knowledge is a hard road, but maybe necessary. - 1 more reply
New conversation -
-
-
@GaryMarcus@karpathy Pragmatic#machineintelligence means less - examples, data, new sources, data science - and more#GuidedLearningThanks. 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.