HetSys News
A review of TOTEMIC Training School 2025: "Tools for Energy Materials Modelling Acceleration", Corsica
HetSys students travel far and wide to attend conferences, workshops and training schools relevant to their PhD. They often give talks and present posters of their research as their projects develop through the four years of their study. George Simmons, a first year HetSys student recently attended the TOTEMIC Training School 2025: "Tools for Energy Materials Modelling Acceleration" in Corsica, France and gives us an insight into his experience below:
The TOTEMIC Training School provided an open and engaging environment for networking and learning about projects across Europe that aim to improve materials discovery for sustainable energy technology. The main technology that was being promoted was the use of machine learning and artificial intelligence to accelerate the development of materials via “Materials Acceleration Platforms (MAPs)”, which aim to automate the materials science design and discovery process, which conventionally takes a matter of weeks or months to complete.
A few highlights that I found particularly interesting were during the talks from Professor Helge Stein and Dr Kevin Rossi. Helge gave an overview of the robotics engineering that goes into MAPs, and showcased networking solutions for linking automated experimental platforms to other scientists around the world. This was particularly intriguing, as it displayed a different technical side to the science itself that was being performed, which was chemical engineering. This also led to interesting discussion on how tools in Python could be used.
Rossi provided an overview of how he utilised various techniques in machine learning to classify the metallic atoms present sitting on a piece of material from electron microscope images. His method utilised convolutional neural networks to identify features in the images, then he encoded the features on a latent data space of lower dimensionality, then used a Gaussian mixture model to classify the atoms. This was interesting as it showcased how various techniques (which are even taught to us in the training module ES98E) can be used to perform scientific procedures.
Overall, the week provided an interesting dive into how different scientists from different backgrounds use artificial intelligence to make informed decisions in their own research.