ľ¹Ï¸£ÀûÓ°ÊÓ

Yesterday, former master's student Rinske Alkemade received the 3rd prize of the KHMW's Young Talent Award. De Koninklijke Hollandsche Maatschappij der Wetenschap (KHMW) presents the Young Talent Award annually to promote scientific education in technical and science subjects. Besides the Young Talent Award, Rinske also won the Thesis prize '22 of the Department of Physics at Utrecht ľ¹Ï¸£ÀûÓ°ÊÓ.

The KHMW jury was very impressed by Rinske's thesis on "the movement of particles in a glassy material" and the fact that it already has been  She wrote this publication together with her thesis supervisors Frank Smallenburg CNRS, France) and Laura Filion (Utrecht ľ¹Ï¸£ÀûÓ°ÊÓ). Laura is also the one who nominated Rinske for the Young Talent Award.

"Rinske is an exceptionally talented and creative physicist. She is careful despite being extremely efficient, and displays impressive levels of independence, initiative, and scientific insight". Laura Filion

Glass, but not a tea glass or a window

Rinske's thesis is about the movement of particles in glassy material. When people think of glass, they often immediately think of a window or a tea glass, but scientifically speaking, there are many more "glass-like" materials. In fact, glass appears solid, but there is actually movement in the particles. In some areas of the glass, the particles move faster, than in other areas. In her thesis, Rinske focused on whether the difference in mobility between different areas can be explained by the way the particles are structured.  

Training algorithms

To investigate the relation between structure and mobility, Rinske used machine learning in her research. Very specifically, this means that she expressed the local structure of each particle in parameters and then trained algorithms to predict the mobility of each particle based on these parameters. The idea behind this is that although we humans may not be able to see how structure affects mobility, we may be able to train algorithms capable of doing so. At the beginning of her research, she compared three different machine learning techniques. From this comparison, the simplest technique proved to be the best and also the easiest to interpret. The latter is important because, after all, the goal is not only to be able to make good predictions, but also to understand what the algorithm bases its predictions on. She then looked at what we can learn from this algorithm about the relationship between structure and mobility.

Follow-up research

In her thesis, Rinske also made a proposal for three follow-up studies, which she is working on herself as a PhD student at Utrecht ľ¹Ï¸£ÀûÓ°ÊÓ. Currently, she is working on determining the newly found insights are universal and therefore also apply to other glass systems than the one she initially looked at.

"What I liked most about writing my thesis was that all the puzzle pieces came together at the end. This allows me to tell a story." Rinske Alkemade

Telling the story

"I learned a lot from my supervisors Frank and Laura, such as thinking in an intuitive, scientific way. This helps you learn what is interesting in your research and which things to let go of. I also learned how to communicate your story well, so that it is understandable to everyone."