Modeling Microwave S-parameters and BO for the optimal operation of a nuclear facility

- 24/04/2025

It was truly an honour to host two speakers today:

Thijs Ullrick - 'Modeling Microwave S-parameters using Frequency-scaled Rational Gaussian Process Kernels'

Bio: A PhD student at Department of Information Technology, University of Ghent, working on Variability aware design and Compact circuit models and extraction.
Abstract: This work presents a machine learning technique to model the complex-valued scattering parameters (S-parameters) of passive microwave devices as a function of frequency and a set of design variables. The proposed Gaussian process (GP) model intricately models the real and imaginary parts of the S-parameters by employing a physics-informed kernel, adept at representing complex holomorphic functions and incorporating the Hermitian symmetry inherent in scattering parameters. Additionally, to extend the kernel's capabilities to higher dimensions beyond standard GP techniques, it is extended with a frequency scaling, enhancing the modeling capacity.

The resulting physics-informed frequency-scaled GP model accurately predicts the S-parameter values at desired parameter configurations in the design space. One application example demonstrates the superiority of the new kernel, compared to standard GP kernels

Santiago Ramos - 'Gaussian process and Bayesian optimization for the optimal operation of a nuclear facility'

Bio: A PhD student at Department of Electromechanical Engineering, University of Antwerp, working on Optimization and control of the Isotope Separator On-Line technique (ISOL).
Abstract: Since its development, the Isotope Separation On-Line (ISOL) technique has driven advancements in various scientific fields, particularly nuclear medicine, by enabling the production of isotopes for cancer treatment and diagnosis. Optimizing ISOL facilities requires complex tuning, traditionally performed by experienced operators. However, this process is time-consuming and often suboptimal due to the many parameters involved. Optimization algorithms have thus emerged as valuable tools for supporting the tuning process. ISOL facility tuning presents challenges such as high dimensionality, constrained design spaces, and noisy observations. This presentation provides insight into ISOL tuning from an optimization perspective using Gaussian process and Bayesian optimization as primary tool to optimize the performance of the ISOL technique.