Higher-order gradients

[ddrous]

Hi,

First of all, thanks for such a wonderful project !

In the ICLR paper attached to this project, you mentioned the need for gradients not only of functionals, but also for operators.

Is there currently support for second- and third order gradients ? If not, is there a plan to support this in the future ? Finally are you welcoming contributions ?

I'm ultimately interest in truncated Taylor polynomials for functionals in order to replace the current parametrisation->vector-valued differentials I'm using.

Thank you,