Zaragoza hosts the international "Machine Learning-Assisted Sampling" school on AI-driven simulation

The increasing integration of artificial intelligence and machine learning (ML) into scientific research is opening new pathways for the analysis and simulation of complex physical systems. In this context, ARAID researcher David Yllanes Mosquera is organizing the CECAM Flagship School on Machine Learning-Assisted Sampling, which will take place from June 15 to 19, 2026, at ZCAM in Zaragoza. The school is aimed at students and early-career researchers and will provide advanced training focused on the application of ML techniques to improve the sampling of complex probability distributions, a key aspect in physical simulations.

Sampling plays a fundamental role in areas such as molecular simulations, where it is used to study interactions between particles and their environment, as well as in complex systems where contact with a thermal bath leads to stochastic dynamics. In quantum systems, whose nature is intrinsically probabilistic, sampling techniques are also essential for accurate simulation. Over the course of the week, the program will cover topics including autoregressive models, learning dynamical laws from trajectories, neural networks for quantum simulations, mathematical learning models for inference, and tensor networks.

The school will feature renowned international experts, including Federico Ricci-Tersenghi (La Sapienza, Rome), Miguel Ruiz-García (UCM), Zakari Denis (Max Planck Institute), Roger Guimerá (Universitat Rovira i Virgili), and Xavier Waintal (INAC CEA Grenoble). This initiative strengthens Zaragoza’s international profile as a hub for advanced scientific training and contributes to the development of innovative methodologies based on artificial intelligence, with significant impact on physics and related research fields.