On August 23-27, 2021 23rd International Conference-School “Advanced Materials and Technologies” was organised in Palanga, Lithuania. This year, the conference attracted the record number of participants: 244 students, researchers, lecturers from 15 countries participated either on-site or online.
Poster presentation of Kaunas University of Technology, Institute of Materials Science PhD student Tomas Klinavičius, titled “Neural Network Enabled Prediction of Power Reflection of Polarized Light from Material Surface”, was awarded the “Young Scientist Award” as selected by an international jury.
According to doctoral student T. Klinavičius, neural networks are machine learning algorithms capable of determining complex relations in data that depends on multiple variables. Optical response of a photonic device can be evaluated by using a multitude of numerical modelling methods. Numerical methods, while accurate, are rather slow. Meanwhile properly trained neural networks can provide a result of similar accuracy much faster. In his poster, the student presented architecture parameter optimization of several different neural networks which allowed to achieve their optimum performance compared with results obtained by numerical modelling.
The conference-school was organised by Institute of Materials Science of Kaunas University of Technology, Lithuanian Materials Research Society, Vilnius University, University of Latvia, Riga Technical University, University of Tartu, Center for Physical Sciences and Technology, University of Southern Denmark, and Applied Research Institute for Prospective Technologies.