summerfieldlab: Neuroscientists are turning to deep networks as computational theories for biology. But can they be used to make falsfiable predictions about neural representation, dynamics, behavior, and cognition? Answers in a new review with @SaxeLab @steph_nelli
3 replies, 485 likes
Andrew Saxe: New perspective paper with @steph_nelli and @summerfieldlab! How can our growing theoretical understanding of deep networks aid our understanding of neural and cognitive phenomena?
1 replies, 105 likes
Alex Naka: If deep learning is the answer, then what is the question?
I'm gonna go with "how do I get people to give me money" https://t.co/ehAsIVIdJH
4 replies, 62 likes
Vincent Costa: This is a really thoughtful review of how neuroscientists should approach use of deep learning as a tool for hypothesis generation/testing; especially regarding how DL can be used to query cognition and motivated behavior.
0 replies, 18 likes
Steph Nelli: If deep learning is the answer, what is the question?
Feedback and thoughts welcome!
0 replies, 17 likes
Esther Mondragón: Interesting! Below a rather personal summary
"sophisticated behaviours and structured neural representations observed in humans and other animals might emerge from a limited set of computational principles"
2 replies, 5 likes
Dan Bang: Looking forward to reading this from the amazing @SaxeLab @steph_nelli and @summerfieldlab
0 replies, 4 likes
Steph Nelli: If deep learning is the answer, then what is the question?
Instead of blindly seeking correlations between brains and “black box” networks, neuroscientists can use deep learning to develop canonical sets of neural theories.
Read more here:
1 replies, 3 likes
Brundage Bot: If deep learning is the answer, then what is the question?. Andrew Saxe, Stephanie Nelli, and Christopher Summerfield http://arxiv.org/abs/2004.07580
1 replies, 0 likes
Found on Apr 17 2020 at https://arxiv.org/pdf/2004.07580.pdf