We are a small interdisciplinary team, working on theoretical and computational methods for quantum mechanical rational compound design (QM-RCD).
Our lab was established in July of 2013.
Our research falls into the realm of physical chemistry, and relies on theory and computations.
Due to the interdisciplinary nature of our work, we rely heavily on physics, mathematics, and computer sciences,
and we are active members of the university's
Competence Center of Computational Sciences.
Unfortunately, we lack the capacity to carry out our own experiments.
Instead we compare our computational predictions to experimental results reported in the scientific literature,
or we collaborate directly with experimentalists.
Aug 2018: PhD position available! Together with the
Nash-group we have just been awarded
funding for a PhD student project by the Swiss Nanoscience Institute!
The project deals with unsupervised learning for directed evolution of anti-adhesive peptides to fight antibiotic resistant bacteria.
It starts on Jan 1 2019, and the Nash group will perform validating AFM experiments. Interested candidates should apply to our group asap!
Jul 2018: Diana's paper was accepted in JCTC. This was a lot of work, Go Diana!
Jun 2018: Samuel got an offer to join Google, congratulations!
Apr 2018: Marco got an offer to join Toyota, congratulations!
Dec 2017: Our CECAM workshop "Machine Learning at Interfaces", scheduled to take place Jun 4-8 at EPFL, has been approved!
Dec 2017: We have been awarded an ERC Consolidator grant!
Jul 2017: OUTREACH: The Science+Fiction Festival 2017 panel discussion, moderated by Bernard Senn (SRF2) and with Anatole and Profs. Zimmerli (HU Berlin) and Riener (ETHZ),
on ``Artificial Intelligence - wann werden wir ueberfluegelt'', can be viewed here https://www.youtube.com/watch?v=6j3t-iqQxfM
Jun 2017: Samuel Chang has become Dr. Chang. Congratulations, Samuel!
Feb 2017: Google collaboration: Our first draft has been posted on the arxiv:
"Fast machine learning models of electronic and energetic properties consistently reach approximation errors better than DFT accuracy",
Jan 2017: The "53rd Symposium on Theoretical Chemistry-Big Data in Chemistry: From Molecular Structure to Condensed Phase Dynamics",
organized in August in Basel by the Meuwly group and us, has come online!
For registration and further details go to www.chemie.unibas.ch/~stc2017.
Jan 2017: Virtual reality immersion in electron density of drug-target complex
Dec 2016: Chemical derivatives of Diels-Alder transition states (generated by Anders)
Nov 2016: Grant awarded "Big Data for Computational Chemistry: Unified machine learning and sparse
grid combination technique for quantum based
molecular design" (NRP 75) by SNF on collaborative project with Harbrecht group in Mathematics
Nov 2016: Mehdi Segidhi joined our group as a Postdoc: Welcome, Mehdi!
Oct 2016: Anders Christensen joined our group as a Postdoc: Welcome, Anders!
Oct 2016: Heini (Stefan) Heinen joined our group as a PhD student: Welcome, Heini!
Sep 2016: Felix Phys Rev Lett on ML energies of 2 million crystals just came out!
Sep 2016: Dirk Bakowies joined our group as a senior research associate: Welcome, Dirk!
Sep 2016: Anne-Sophie Alingue joined our group as a PhD student: Welcome, Anne-Sophie!
Aug 2016: Marco Di Gennaro joined our group as a Postdoc: Welcome, Marco!
Aug 2016: Google grant awarded for machine learning!
Jun 2016: Anatole accepted an offer to return to the University of Basel.
Oct 2015: Rafael Sarmiento Perez joined our group as a Postdoc: Welcome, Rafael!
Sep 2015: Bing Huang joined our group as a Postdoc: Welcome, Bing!
Aug 2015: Our paper on ML predictions of atomic properties is out! Have a look at an atom in a molecule:
Moving vertical red bar indicates instantaneous C-NMR shift of atom highlighted in yellow.
The bar advances at a rate of 8 molecules per second, as the atom navigates through different molecular environments
drawn from a data-set of over 50 k C-NMR shifts whose distribution is given on the horizontal axis.
See here for more details:
"Machine Learning for Quantum Mechanical Properties of Atoms in Molecules", M. Rupp, R. Ramakrishnan, OAvL,
J. Phys. Chem. Lett.6 3309 (2015).arxiv.org/abs/1505.00350
Apr 2015: Grant awarded for machine learning of reaction rates!
Mar 2015: Anatole's talk on "Machine Learning Models in Chemical Space" in February at UCLA came
Jan 2015: Zhenwei Li joined our group as a Postdoc. Welcome, Zhenwei!
Jan 2015: Diana Tahchieva joined our group as a PhD student. Welcome, Diana!
Nov 2014: Felix Faber joined our group as a PhD student. Welcome, Felix!
Dec 2013: The Swiss National Science Foundation awarded 18M CHF to our
MARVEL proposal (lead: Prof. N. Marzari at EPFL)
on computational materials design.
Nov 2013: The U.S. Department of Energy's Office of Science awarded compute time on the BG/Q computer Mira, No 5 on November 2013 top500 list:
125M core hours to be used for ``SiO2 Fracture: Chemomechanics with a Machine Learning Hybrid QM/MM Scheme''within our collaboration with
James Kermode and Alessandro de Vita (King's College, London) at Argonne's Leadership Computing Facility.
56M core hours for our collaboration on ``Non-Covalent Bonding in Complex Molecular Systems with Quantum Monte Carlo'' with Dario Alfe, Mike Gillan, Angelos Michaelides, Mike Towler (All University College London), Ken Jordan (University of Pittsburgh) and Alexandre Tkatchenko (Fritz-Haber Institute, Berlin).
announcement, and the complete
list of awards made.
The goal of our work is to develop and apply algorithms meant to enable the rational design of new materials from first principles,
i.e. the virtual identification of interesting materials/catalysts/molecules using ``smart'' optimizations, rather than mere trial and error.
Once this is accomplished, these algorithms could be combined within new software, such as the materials project's
Lithium battery explorer for designing better electrodes.
molecules, liquids, and solids including molecular crystals and defects
Our calculations are mostly based on combined statistical mechanics and electronic structure
theory software, such as CPMD, SeqQuest, VASP, abinit, octopus, LAMMPS, QUANTUM ESPRESSO, and others.