"Despite years of work we have failed to find cars that can be considered robust on stairs and mountains from the point of human legged locomotion". Let's abandon the whole idea of wheels, then.
“despite years of work we have failed to find neural network models that can be considered robust on MNIST from the point of human perception." important conclusion from @bethgelab, h/t @dileeplearning https://arxiv.org/abs/1805.09190
-
-
- 2 more replies
New conversation -
-
-
Analysis by synthesis played a central role in a lot of models developed in the
@IARPAnews MICrONS program. I am still a strong believer in this theory of cortical computation.@MatthiasBethge is this paper derivative from your MICrONS work? -
@RJVneurotech Indeed this paper is heavily inspired by our work in the MICrONS program and days of discussions as to what the role of the ubiquitous cortical feedback connections could potentially be. Funding sources will be added to the paper next week along with more results.
End of conversation
New conversation -
-
-
Requires sensory motor integration (saccades)
-
Related to this paper: https://arxiv.org/pdf/1511.06292.pdf …
End of conversation
New conversation -
-
-
Or may be human brain may look like neural network physically but in reality it has an virtual abstraction layer over it that is not visible physically and does not look anything like a neural network. Imaging trying to discover TCP packet structure using current measurements.
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.