Machine learning potentials, powered by state-of-the-art neural networks, are used to predict the forces and energies from hour-long quantum calculations in milliseconds. In principle, such an approach could reveal the real-time atomic dynamics of everything from materials to drugs binding on proteins. However, their parameterization is slow, with the construction of good training sets often costing years of a PhD. In this presentation, I will talk about how we at the Hylleraas Centre for Quantum Molecular Sciences developed an active learning approach to train quickly robust models for interactions between atoms.
Program
11:30 – Doors open and lunch is served
12:00 – "Learning forces and energies of atoms by active learning" by Sigbj?rn L?land Bore (Postdoctoral Fellow, Hylleraas Centre for Quantum Molecular Sciences)
This event is open for all students, PhD candidates, postdocs, and everyone else who is interested in the topic. No registration needed.
About the seminar series
Once a month, dScience will invite you to join us for lunch and professional talks at the Science Library. In addition to these, we will serve lunch in our lounge in Kristine Bonnevies house every Thursday. Due to limited space (40 people), this will be first come, first served. See how to find us here.
Our lounge can also be booked by PhDs and Postdocs on a regular basis, whether it is for a meeting or just to hang out – we have fresh coffee all day long!