Artificial Intelligence for Materials Sciences
Welcome to the homepage of the AiMat (Artificial Intelligence for Materials Sciences) group at KIT in which we work on the development of AI and machine learning methods focusing on their application to materials science questions.
The research group headed by junior professor Pascal Friederich was established in 2020 and has been growing ever since. We are therefore always looking for talented students and researchers in computer or natural sciences to join our team!
Our main research areas are
- Data-driven property prediction and design of materials and molecules
- Accelerated, machine learning based atomistic simulations of materials and molecular systems
- Machine learning methods for autonomous data analysis and decision-making in self-driving labs
The Summer School will be held on 8-12 September at KIT. It will cover multiple applications of machine learning in materials research, e.g. materials design and property prediction, ML-potentials for atomistic simulations, self-driving labs, and materials synthesis predictions.
Website & Application
In the summer term 2025 we offer the lecture "Machine learning for natural sciences" together with the related exercises, as well as a seminar on "Critical topics in AI".
Check them now!
Together with our cooperation partners, we published a closed-loop workflow for the discovery of new organic molecules to increase the efficiency of perovskite solar cells. Our strategy combines high-throughput synthesis of organic semiconductors to create large datasets with Bayesian optimization methods and can be transferred to other areas of materials research, such as the search for new battery materials.
Read the KIT news
We are searching for talented and motivated young researchers with backgrounds in computer science or natural sciences to join the team! Currently available are two PhD projects and one postdoc position. Interested? Check out the job openings!
Job openings
On April 25th, 10 schoolgirls visited our group to try out our new VR application. The use of VR goggles and haptic gloves allowed them to not only see but also “touch” molecules and thus to interactively explore the fascinating world of molecules from a completely new perspective.
More information
Lately, everyone has been talking about AI. But what influence do AI and machine learning methods have in the natural sciences? Will they have a lasting impact on scientific research or will the hype create a bubble that will eventually burst? As part of our CZS Summer School, we would like to discuss these and other questions with our speakers and then engage in conversation with the audience in an informal get-together.
See the KIT event calendar entry here
Today, the AiMat group recognizes the great science done by doctoral researchers and students: Marlen Neubert, Yuri Koide, Yumeng Zhang, Mariana Petrova, Annika Leinweber, Laura Ruple, Klara Eckhardt, Aleksandra Hryncyszyn, Michelle Walter, and many more over the previous years. Thanks also to Stephanie Wolf for providing great support. We are looking forward to welcoming more women into the group in the coming years!
Learn more
Congratulations to Jonas Teufel for the Best Student Paper Award at the xAI World Conference 2023!
Jonas has given a talk entitled "MEGAN: Multi-Explanation Graph Attention Network" in which he presented the results of our recent homonymous arXiv preprint.
The GC-MAC Summer School 2023 will be held on 18-22 September at KIT. It will cover multiple aspects of materials acceleration platforms (MAPs) for energy materials, e.g. automated synthesis and characterization, integration of simulation methods in materials design, and ML methods for MAPs.
Website & Registration