everybody interested in machine learning should read the abstract to this paper. h/t @sandyasmhttps://twitter.com/sandyasm/status/1200403581406285825 …
-
David Hand is a well-known critique of ROC curves. He is not saying that classifier technology is bad but the measure we use to assess them. See H-measure: https://www.hmeasure.net
maybe so, but the note of humility he strikes is just right. when AlexNet recognizes an elephant, it recognize the texture rather than the interrelation between parts. right answer, arguably wrong reasons.
-
-
Deep learning systems could recognise different scales in intermediate layers. True, there are no reasoning capabilities in current classification technologies, though.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Note that, I don't think "we" know how to quantify or build a formal system for reasoning in an automated fashion from scratch. Yes, there are Bayesian Networks or other proof systems reasoning on a graph-structure. Catch: They work on human build graphs-structures.
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.