@NandoDF @hardmaru @goodfellow_ian @Miles_Brundage @fchollet what would be a good way to argue against the "black box" character of DL?
-
In fact, I'm referring to these conditions, something that an outsider scientist can understand
There are many ways to attribute NN predictions back to specific input features. Issue is that these features may not meaningful for humans.
-
-
In the case of visual inputs however this is generally a well solved problem (albeit still an area of research).
-
So, you mean that NN capture the manifold of suitable input-output, right? But I've received complains that this mapping is black magic.
End of conversation
New conversation -
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.