2-week MARVEL-ICTP College: Materials simulations in the age of AI

Jun 01, 2026, 9:00 until Jun 12, 2026, 17:00, ICTP, Trieste, Italy

The 2-week MARVEL-ICTP College in computational materials science will take place on 1-12 June, 2026 at ICTP in Trieste. This college is dedicated first and foremost to the MARVEL community, as well as junior researchers all around the world, with support also for those coming from emerging economies. It aims to bring together graduate students and early-career researchers from all over the world to learn and explore the state-of-the-art in computational materials science in the era of AI, and to forge the next generation of researchers in the field. Registrations are closed.

The two-week programme includes lectures and hands-on sessions on electronic-structure and atomistic simulations to understand, predict, and design material properties. It includes keynote talks and lectures by distinguished scientists (see poster attached) and covers first-principles methods, molecular dynamics simulations, and machine-learning approaches. The programme will also feature two poster sessions, one in each week, where participants can present and discuss their work, and open-mic sessions to brainstorm together.

Week 1 focuses on electronic-structure theory and spectroscopies (density-functional theory and density-functional perturbation theory, many-body perturbation theory, topology and Wannier functions, advanced functionals).

Week 2 covers statistical mechanics and atomistic simulation, including enhanced sampling and machine-learning interatomic potentials. Hands-on labs deploy open source community codes, with poster sessions and open-mic discussions.

Topics

  • Density-functional theory (DFT) and beyond
  • Linear response and lattice dynamics (DFPT, phonons)
  • Electronic, vibrational and core-level spectroscopies
  • Many-body perturbation theory (GW, BSE)
  • Topology and polarization (Berry phases, Wannier functions)
  • Advanced functionals (Hubbard, Koopmans)
  • Introduction to statistical sampling by molecular dynamics
  • Sampling of different ensembles (thermostats, barostats, replica exchange)
  • Rare events and enhanced sampling (metadynamics, accelerated MD)
  • Machine learning for materials (descriptors, potentials, training and validation)
  • Advanced ML for sampling (fine tuning, direct force prediction, force-free integration)

Speakers
C. Clementi, Free University of Berlin, Germany
C. Dellago, University of Vienna, Austria
G. Galli, University of Chicago, USA
G. Kresse, University of Vienna, Austria
M. Rossi, University of Cambridge, UK

Lecturers
S. Baroni, SISSA, Italy
G. Bussi, SISSA, Italy
M. Calandra, University of Trento, Italy
M. Cococcioni, University of Pavia, Italy
N. Colonna, Paul Scherrer Institute, Switzerland
S. De Gironcoli, SISSA, Italy
P. Delugas, SISSA, Italy
A. Ferretti, CNR, Italy
R. Gebauer, ICTP, Italy
P. Giannozzi, University of Udine, Italy
A. Hassanali, ICTP, Italy
A. Laio, SISSA, Italy
E. Linscott, Paul Scherrer Institute, Switzerland
A. Marrazzo, SISSA, Italy
L. Monacelli, Sapienza University of Rome, Italy
S. Narasimhan, JNCASR, India
M. Palummo, Tor Vergata University of Rome, Italy
R. Resta, University of Trieste, Italy
I. Souza, University of the Basque Country, Spain
I. Timrov, Paul Scherrer Institute, Switzerland
D. Varsano, CNR, Italy

Organizers
Sara Bonella (CECAM, EPFL, Switzerland)
Michele Ceriotti (EPFL, Switzerland)
Nicola Marzari (EPFL, Switzerland)
Sandro Scandolo (ICTP, Italy)