Bayesian optimization with improved scalability and derivative information for efficient design of nanophotonic structures

TL;DR

This paper introduces an enhanced Bayesian optimization method that leverages shape derivatives and iterative inversion to efficiently design nanophotonic structures, especially when many iterations and derivative data are involved.

Contribution

It combines forward shape derivatives with iterative inversion in Bayesian optimization, enabling scalable and derivative-informed design of nanophotonic devices.

Findings

Successfully optimized a waveguide edge coupler

Demonstrated improved scalability with derivative information

Extended Bayesian optimization applicability to complex designs

Abstract

We propose the combination of forward shape derivatives and the use of an iterative inversion scheme for Bayesian optimization to find optimal designs of nanophotonic devices. This approach widens the range of applicability of Bayesian optmization to situations where a larger number of iterations is required and where derivative information is available. This was previously impractical because the computational efforts required to identify the next evaluation point in the parameter space became much larger than the actual evaluation of the objective function. We demonstrate an implementation of the method by optimizing a waveguide edge coupler.