Exact adjoint sensitivity analysis for neural-based microwave modeling and design

TLDR

The study introduces an adjoint neural network method for sensitivity analysis in neural-based microwave modeling and design. The method trains original and adjoint neural networks simultaneously, enabling efficient first- and second-order sensitivity analysis across generic microwave neural models that embed empirical data. The technique facilitates neural-based microwave optimization, unified device modeling, and circuit design, as shown by demonstrations on VLSI interconnects, large-signal FETs, and three-stage power amplifiers.

Abstract

For the first time, an adjoint neural network method is introduced for sensitivity analysis in neural-based microwave modeling and design. The proposed method is applicable to generic microwave neural models including a variety of knowledge-based neural models embedding microwave empirical information. Through the proposed technique, efficient first- and second-order sensitivity analysis can be carried out within the microwave neural network infrastructure using neuron responses in both the original and adjoint neural models. A new formulation of simultaneous training of original and adjoint neural models allows robust model development by learning not only the input/output behavior of the modeling problem, but also its derivative data. The proposed technique is very useful for neural-based microwave optimization and synthesis, and for analytically unified DC/small-signal/large-signal device modeling and circuit design. Examples of high-speed very large scale integration system interconnect modeling and optimization, large-signal FET modeling, and three-stage power-amplifier simulation utilizing the proposed sensitivity technique are demonstrated.