Title: Neural collapse with unconstrained features
Abstract: Neural collapse is an emergent phenomenon in deep learning that was recently discovered by Papyan, Han and Donoho. We propose a simple “unconstrained features model” in which neural collapse also emerges empirically. By studying this model, we provide some explanation for the emergence of neural collapse in terms of the landscape of empirical risk. (Joint work with Hans Parshall and Jianzong Pi.)
Here is the new link and meeting ID+passcode:
https://wse.zoom.us/j/91467375713?pwd=VjN3ekZTRFZIWS80NnpwZUFRUzRWUT09
Meeting ID: 914 6737 5713
Passcode: 272254