Training of Continuum Models

The current dominant training paradigm is back propagation, e.g. SGD or its variants. The continuum perspective provides avenues for developing new training formulations beyond backpropagation. We shall focus on the continuous formulation and discretisation (connecting with WP 1.2), and their discrete analogues (connecting with WP 2.3) and in particular will look at score based diffusion.

Get in Touch!

To subscribe to the mailing list

Send an email request