Papers I'm thinking about lately (all of my papers)
-
The Platonic Representation Hypothesis
Minyoung Huh*, Brian Cheung*, Tongzhou Wang*, Phillip Isola*, ICML 2024 Oral
TLDR: As all models continue to improve, the representations they use are converging.
-
System Identification of Neural Systems: If We Got It Right, Would We Know?
Yena Han, Tomaso Poggio, Brian Cheung, ICML 2023
TLDR: Even if we had a perfect model of the brain, we probably wouldn't know it.
-
Superposition of many models into one
Brian Cheung, Alex Terekhov, Yubei Chen, Pulkit Agrawal, Bruno Olshausen, Neurips 2019
TLDR: You can literally add a bunch of model parameters together and still get good performance.
-
Emergence of foveal image sampling from learning to attend in visual scenes
Brian Cheung, Eric Weiss, Bruno Olshausen, ICLR 2017
TLDR: If you simulate the evolution of a retina with a goal to attend, a fovea emerges.