Concept of continuous depths for neural networks in UDEs

I don’t think that framework of thinking even makes sense with UDEs. Take a look at this recent talk:

It’s all about mechanistic models. If you’re not talking mechanistic models, then you’re not talking UDEs.

1 Like