Neural SDE vs Neural ODE

May I please ask a question, please redirect me when this is the wrong place to ask :slight_smile:

What is the main (dis)advantage of neural SDEs over neural ODEs?
Can they somehow tackle ‘more difficult’ problems or somehow provide more information?

Thank you!

Forget about neural part for now. SDEs and ODEs are just completely different class of problems. It not like one is better or worse than the other, they model different behaviours.


SDEs are much more general. As always, anything that’s more general is going to be much more computationally difficult to compute. They include essentially an uncertainty estimate through the noise term, but you have to pay for that.

1 Like