News and Seminars

© 2020 EPFL

Artificial intelligence explains hydrogen's behavior on giant planets

— Using computer simulations powered by machine-learning algorithms EPFL scientists have made an important breakthrough in understanding how hydrogen behaves on Saturn and Jupiter. Their research has just been published in Nature.

© COSMO / 2019 EPFL

Barely porous organic cages for hydrogen isotope separation

— Porous materials obtained by assembling cage-shaped molecules of different types achieve high selectivity in separating hydrogen and deuterium gas by means of a quantum sieving process.

© 2019 Stuart Althorpe

Prof. Althorpe gives a seminar on quantum dynamics

— Prof. Stuart Althorpe (Cambridge University), who is visiting the Laboratory of Computational Science and Modelling, will give a seminar on "Real-time dynamics from imaginary-time path-integrals: theory and practice" on Tuesday October 29, 16:00, Room MED 2 1124 (CoViz2)

A. Fabrizio © 2019 EPFL

Machine learning the electron density

— A collaboration between the Laboratory of Computational Science and Modelling and the Laboratory for Computational Molecular Design developed a transferable and scalable machine-learning model capable of predicting the total electron density directly from the atomic coordinates. This model has been trained on a database of small molecular fragment, and used to predict the electronic density of polypeptides, and to compute electrostatic interactions between fragments, that contribute to the stability of proteins and the interactions between enzymes and drug molecules.

Prof. Joel Bowman awardin the poster prize to Felix © 2019 EPFL

Best poster award for Félix Musil

— Félix Musil won a poster prize at the workshop "Developing High-Dimensional Potential Energy Surfaces – From the Gas Phase to Materials" in Göttingen, for his poster on "Atom-density representations for machine learning". Congratulations!

© 2019 EPFL

Towards the design of molecular materials with Dr. Brandenburg

— On January 24 at 17:00, MED 2 1124 (CoViz2), Dr. Jan Gerit Brandenburg will present his work on crystal structure prediction, discussing both the challenge of obtaining very accurate estimates of polymorph energetics, and that of exploring the enormous search space of possible stable configurations.

Machine learning accelerates quantum mechanical modelling of water © COSMO / 2019 EPFL

Machine learning and quantum mechanics team up to understand water

— Why is water densest at around 4 degrees Celsius? Why does ice float? Why heavy water has a different melting point compared to normal water? Why do snowflakes have a six-fold symmetry? A collaborative study, led by researchers in EPFL and just published in the Proceedings of the National Academy of Sciences, provides physical insights into these questions by marrying data-driven machine learning techniques and quantum mechanics.

A data-driven construction of the periodic table of the elements © 2019 EPFL

A data-driven construction of the periodic table of the elements

— Researchers in the Laboratory of Computational Science and Modeling at EPFL show how the grouping of elements that underlies the periodic table of the elements arises from an automatic analysis of atomistic simulation data, that also improves substantially the performance of machine-learning models that make it possible to sidestep expensive quantum mechanical simulations.

©Michele Ceriotti / 2018 EPFL

AI and NMR spectroscopy determine atoms configuration in record time

— EPFL scientists have developed a machine-learning approach that can be combined with experiments to determine, in record time, the location of atoms in powdered solids. Their method can be applied to complex molecules containing thousands of atoms and could be of particular interest to the pharmaceutical industry.

ML Accuracy as a function of descriptor size © COSMO / 2018 EPFL

Automatic Selection of Descriptors for Machine Learning of Materials

— What is the most effective representation of atomistic models to be used as the input for machine learning? A paper from the Laboratory of Computational Science and Modelling demonstrates how to let an algorithm make the optimal choice.

All news