Derivative-Informed Neural Operator: An efficient framework for high-dimensional parametric derivative learning

Thomas O'Leary-Roseberry, Peng Chen, Umberto Villa, Omar Ghattas

Abstract: Highlights•Adding derivative information to neural operator training improves function accuracy.•Neural operators trained without derivatives yield inaccurate derivative predictions.•Reduced basis neural operators lead to efficient derivative learning formulations.•DINOs efficiently and accurately approximate derivatives used in outer-loop problems.