Machine Learning Interatomic Potentials: Theory and Practice
Machine learning interatomic potentials (ML-IPs) have now established themselves as a key technique in atomistic modelling. They allow the simulation of many diverse types of systems, from molecular to the solid state, at the accuracy of highly sophisticated electronic structure methods but at a greatly reduced cost [1-8]. Applications of ML-IPs to real, complex scientific problems continue to appear at an accelerating pace, and it is clear they are becoming a key technique needed to tackle atomistic simulation problems with broad potential societal benefit – including clean energy generation and storage systems , new structural and functional materials [7,8], and modelling of biomolecular systems with potential to improve our fundamental understanding of biochemical processes and apply this understanding to the treatment of many important diseases .
In order to fully realize the potential of ML-IPs to make progress on these challenges, there is a pressing need to educate young and early-career researchers in the practical details of fitting and using machine learning potentials in research, especially including advanced or emerging techniques such as fitting vector-valued  or tensorial  physical quantities coupled with the ML-PES as well as incorporating long-range  and equivariant  physics into the fit itself. In addition, we aim to provide researchers who have promising research ideas that could benefit from the application of ML-IPs a pathway to enter the field and make valuable contributions, not only in the application but also in the theory of ML-IPs. We envision this School as being especially helpful to those researchers who would otherwise find it difficult to enter the field because of a lack of expertise or connections in their local network or a lack of funding to attend important international conferences in this rapidly evolving field.
Update 26.08.2023: Applications are now closed. Thank you to everyone who applied; we will be sending out all remaining decisions in the next few days.
Miguel Caro (Aalto University) - Organiser
Max Veit (Aalto University) - Organiser
Felix-Cosmin Mocanu (École normale supérieure) - Organiser
Sanggyu Chong (EPFL) - Organiser
Federico Grasselli (EPFL) - Organiser
Kevin Kazuki Huguenin-Dumittan (EPFL, Lausanne, Switzerland) - Organiser
Jigyasa Nigam (Ecole Polytechnique Federale de Lausanne) - Organiser
Davide Tisi (EPFL) - Organiser
Chiheb Ben Mahmoud (University of Oxford) - Organiser
Carlo Maino (University of Warwick) - Organiser