Diffusion Model Inverse Modeling and Applications to Microwave Filters
Open Access
- 26 January 2026
- journal article
- research article
- Published by MDPI AG in Electronics
- Vol. 15 (3) , 527
Abstract
This paper presents a framework for inverse modeling of microwave filters based on a conditional diffusion model developed to address the intrinsic non-uniqueness of reconstructing coupling matrices from specified S-parameter responses. In the forward diffusion process, Gaussian noise is progressively added to the filter design variables, and a denoising network conditioned on the target electrical responses is trained to predict the injected noise at arbitrary diffusion steps. At inference, we initialize with Gaussian noise and execute the learned reverse denoising dynamics process; independent seeds yield diverse sets of physically feasible design-variable solutions that satisfy identical electrical-response constraints. Experiments on fourth- and sixth-order filters show that the proposed method outperforms multivalued neural networks (MVNNs) and conditional generative adversarial networks (CGANs) in prediction accuracy, solution diversity, and cumulative training cost, thereby providing a robust and efficient framework for inverse microwave-filter modeling.Keywords
Funding Information
- National University of Defense Technology (BHCX-2023060101)
- Tianjin Municipal Science and Technology Program (23ZGZNGX00040)
This publication has 24 references indexed in Scilit:
- Inverse Design of Diffractive Metasurfaces Using Diffusion ModelsACS Photonics, 2025
- Degradation Parameters Estimation of Induction Motors Using Hybrid-Driven MethodIEEE Transactions on Instrumentation and Measurement, 2025
- Development, Characterization, and Circuit Modeling of Inkjet-Printed Coupled Ring Resonators for Application in Biological SamplesIEEE Transactions on Instrumentation and Measurement, 2023
- A Supervised Surrogate-Assisted Evolutionary Algorithm for Complex Optimization ProblemsIEEE Transactions on Instrumentation and Measurement, 2023
- Methods to Expand the Bandwidth of the IC-Stripline CellIEEE Transactions on Instrumentation and Measurement, 2022
- Conditional Generative Adversarial Networks for Inverse Design of Multifunctional MetasurfacesAdvanced Photonics Research, 2022
- Inverse Modeling for Filters Using a Regularized Deep Neural Network ApproachIEEE Microwave and Wireless Components Letters, 2020
- Low-Cost Surrogate Models for Microwave FiltersIEEE Microwave and Wireless Components Letters, 2016
- Simple and Efficient High-Dimensional Parametric Modeling for Microwave Cavity Filters Using Modular Neural NetworkIEEE Microwave and Wireless Components Letters, 2011
- Neural networks for microwave modeling: Model development issues and nonlinear modeling techniquesInternational Journal of RF and Microwave Computer-Aided Engineering, 2000