Fraunhofer Institute for Integrated Systems and Device Technology IISB
The Computational Lithography and Optics Group warmly congratulates Vlad Medvedev on successfully defending his PhD thesis titled “Physics-Informed Machine Learning for Modeling and Design of Nano-Optical Devices" on July 1, 2026.
In his doctoral research, Vlad addressed a key challenge in modern computational optics: how to model light interaction with nano-optical devices accurately and efficiently. Structures such as EUV masks and optical metasurfaces have features on the scale of the wavelength of light, making their simulation highly demanding with conventional electromagnetic solvers, especially in complex three-dimensional cases. To overcome this, Vlad explored physics-informed machine learning, combining neural networks with the fundamental laws of optics and electromagnetism. This approach aims to keep the speed of machine learning while improving physical reliability and reducing the need for large training datasets.
His thesis focused on three main areas: the simulation of light diffraction from EUV masks, including 3D mask and imaging effects, as well as the forward modeling and inverse design of optical metasurfaces. Through this work, Vlad showed that physics-informed machine learning can provide fast and reliable surrogate models for rigorous simulations and can also support efficient device design and optimization. More broadly, his research highlights the strong potential of physics-informed machine learning to become an important tool in routine nano-optical simulation and design, helping to make computational optics workflows faster, more scalable, and more practical.
We sincerely congratulate Vlad on this impressive achievement and thank him for his dedication, scientific curiosity, and valuable contributions to the group. We wish him every success in the exciting journey ahead!