Šimon Kratochvíl, a student at Brno-Řečkovice Grammar School, achieved remarkable success in 2025 at the nationwide finals of the Secondary School Professional Activity competition (SOČ), placing 4th in the physics category. His project, supervised by Pavel Ondračka from the Department of Plasma Physics and Technology at our faculty, focused on leveraging open data and machine learning to predict new materials.
How would you briefly describe the topic of your Secondary School Professional Activity competition (SOČ) project? Who supervised it?
My project is titled Using Open Data for Predicting New Materials with the Help of Machine Learning. It was supervised by Dr. Pavel Ondračka from the Department of Plasma Physics and Technology. At our grammar school, students must complete a year-end project in the penultimate year. I had a sense I wanted to explore a topic bridging physics and computer science. This specific subject, which I found on Masaryk University’s website, caught my attention because it so nicely connects the two fields.
How long have you worked on your project, and what did it involve?
I have been working on the project for over a year, starting in March 2024. Throughout, I needed to learn certain physics principles related to materials and also develop programming skills in Python, which I used to train machine learning models.
How does training machine learning models work in this area?
Nowadays, researchers often create their own training datasets to develop models. The advantage is the reliability of such data, but the downside is that it is extremely time-consuming. My work is unique because it uses freely available materials simulation data, enabling more efficient modeling without the need to generate original datasets. The challenge was to verify whether these open data were sufficiently accurate and usable.
What are open data of materials simulations?
Open data of materials simulations are digital data describing the properties, behavior, and responses of various materials—obtained through simulations or experiments—which are made publicly available without major restrictions. This means that anyone can freely use, analyze, distribute, and repeatedly use such data, for example, in research, education, new product development, or simulation modeling in industry and academia.
What is the practical approach? How do you model materials using machine learning?
For example, you take around ten thousand structures with various crystal lattices. The model learns how atoms interact within these structures. The resulting model can then simulate atomic behavior in space—for instance, what happens when you heat a material and observe how its structure changes. This is an approximate method—not as precise as the Schrödinger wave equation (which is exact by definition but practically unusable for reasonably large systems)—yet it requires far less computing power while remaining accurate enough for many applications.
What are some practical applications of this approach?
Materials modeling has a wide range of applications, from simulating viruses, like the spike protein of SARS-CoV-2, to traditional materials science. A key advantage is that you can model expensive or hazardous materials without physically handling them in a lab, saving both time and resources. My project expands modeling possibilities by introducing a novel approach through using open data.
How was the experience of working with your supervisor?
Working with Pavel Ondračka was excellent. We planned the work systematically together, so I never felt lost or unsure about my next steps.
What did you gain from working on this project?
Most importantly, I had the opportunity to experience real scientific research. Writing a scientific project already in high school has definitely made future work, whether a bachelor’s thesis or other, much easier. As I mentioned, I learned the basics of Python programming, which is broadly useful across natural sciences. Additionally, I now have a wider understanding of what can be studied in physics.
Do you plan to continue studying physics?
Yes, I definitely want to study physics further. I’m keeping my options open regarding specializations because many areas interest me, but physics remains the clear choice.
Would you recommend other high school students to get involved in SOČ?
Absolutely. I recommend it to anyone interested in science in any form.
Thank you for the interview, and we wish you much success in your future studies and research!
At first glance, public transport in Brno may seem barrier-free, but for people with hearing or visual impairments, it still has many obstacles.
After a year, we are once again delighted by the flowers of Victoria, the largest lily in the world. Its enormous leaves, measuring one and a half meters in diameter (up to two and a half meters in nature), and the nocturnal flowers, which change color from white to pink over two nights, attract the...
