BCCMS  /  Events  /  2019  /  CECAM OUTBOX  /  Program
English  /  Deutsch

Program of the International CECAM Workshop

Thinking outside the box - beyond machine learning for quantum chemistry

Conference site:  House of Science, Downtown

Monday, October 7th 2019 (Radisson Blu Hotel Bremen)

18:00

21:00

 

Registration

Tuesday, October 8th 2019 (House of Science Bremen, Downtown)

08:00

08:50

 

Registration

08:50

09:00

 

Opening and welcome, Thomas Frauenheim

Session:

 

Machine learning for Complex Quantum Systems

 

 

Chair: Sheng Meng

09:00

09:40

 

Anatole von Lilienfeld, University of Basel, Switzerland
Quantum machine learning

09:40

10:20

 

Tristan Bereau, Max Planck Institute for Polymer Research,
Mainz, Germany
Modeling intermolecular interactions with physics and ML

10:20

10:50

 

Coffee Break

10:50

11:30

 

Karsten Reuter, Munich University of Technology, Germany
Knowledge-based approaches in catalysis and energy modelling

11:30

12:10

 

Hiromi Nakai, Waseda University, Shinjuku, Japan
Semi-local machine-learned kinetic energy density functional
with third-order gradients of electron density

12:1012:15 
Group Photo

12:15

13:50

 

Lunch Break (Restaurant Q1) and Coffee

13:50

14:30

 

Guanhua Chen, University of Hong Kong, China
Deep learnt exchange-correlation potential

Session:

 

Machine Learning for Structure Prediction

 

 

 

 

Chair: Thomas A. Niehaus

14:30

15:10

 

Bjørk Hammer, Aarhus University, Denmark
Speeding up atomistic structure search with machine learning

15:10

15:50

 

Rickard Armiento, Lynköping University, Sweden
Machine Learning for materials stability

15:50

16:20

 

Coffee Break

16:20

17:00

 

Stefano Leoni, University of Cardiff, UK
ML in multiple timescale molecular dynamics simulations

17:00

17:40

 

Jacek Jakowski, Oak Ridge National Laboratory, Tennessee, USA
Directed transformations of nanomaterials and beam-matter
interactions

19:00

21:30

 

  Welcome Reception (Bremen Town Hall)

Wednesday, October 9th 2019 (House of Science Bremen, Downtown)

Session:

 

Machine Learning for DFTB repulsive interactions (I)

 

 

 

 

Chair: Malte Schüler

08:30

09:10

 

Maxime Van den Bossche, Sorbonne University, Paris, France
Accelerating global optimization searches with an adaptive DFTB parametrization scheme

09:10

09:50

 

David J. Yaron, Carnegie Mellon University, Pittsburgh,
Pennsylvania, USA
A DFTB layer for deep learning of electronic hamiltonians

09:50

10:30

 

Nir Goldman, Lawrence Livermore National Laboratory,
California, USA
Combining ML approaches with DFTB for simulations of reactive
materials

10:30

11:00

 

Coffee Break

11:00

11:40

 

Benjamin Hourahine, University of Strathclyde, Glasgow, UK
Learning around the DFTB model

11:40

-

12:20

 

Stefan Grimme, University of Bonn, Germany
New tight-binding quantum chemistry methods

12:20

14:00

 

Lunch Break (Restaurant Q1) and Coffee

Session:

 

Machine Learning and MD

 

 

 

 

Chair: Adam McSloy

14:00

14:40

 

Kipton Barros, Los Alamos National Laboratory, New Mexico, USA Advances in machine learned potentials for molecular dynamics simulation

14:40

15:20

 

Weitao Yang, Duke University Durham, North Carolina, USA
Machine learning in simulations and force fields with quantum mechanics/molecular mechanics and in DFT

15:20

16:00

 

Stephan Irle, Oak Ridge National Laboratory, Tennessee, USA
Neural network corrected DFTB/MD simulations of long-timescale
self-assembly and transport processes

16:00

16:30

 

Coffee Break

16:30

17:10

 

Roland Mitric, University of Wuerzburg, Germany
Simulation of light-induced nonadiabatic dynamics in molecular aggregates

17:10

17:50

 

Franco P. Bonafé, Max Planck Institute for the Structure and Dynamics
of Matter, Hamburg, Germany
Simulations of impulsive vibrational spectra using Ehrenfest real-time TDDFTB

19:00

22:30

 

Conference Dinner (Restaurant Juergenshof)

Thursday, October 10th 2019 (House of Science Bremen, Downtown)

 

Session:

 

Machine Learning for Quantum Chemistry & Electronic Structure

 

 

 

 

 

Chair: Alessandro Pecchia

 

08:30

09:10

 

Benjamin T. Nebgen, Los Alamos National Laboratory, New Mexico, USA
Hückel theory resurrected: dynamic parameterization of effective hamiltonians using deep learning

 

09:1010:50 
Volker W. Blum, Duke University, Durham, North Carolina, USA
The ELSI infrastructure
 
 

09:50

10:30

 

Ursula Röthlisberger, Swiss Federal Institute of Technology,
Lausanne, Switzerland
Computational Chemistry Meets Artificial Intelligence

 

10:30

11:00

 

Coffee Break

 

11:00

11:40

 

Julian Gebhardt, Fraunhofer Institute for Mechanics of Materials, Freiburg, Germany
Big data approach for next level hybrid perovskite solar cells

 

11:40

12:20

 

Alexandre Tkatchenko, University of Luxembourg, Luxembourg
Towards exact molecular dynamics simulations with quantum chemistry and machine learning

 

12:20

14:00

 

Lunch Break (Restaurant Q1) and Coffee

 

Session:

 

Machine Learning for DFTB repulsive interactions (II)

 

 

 

 

 

Chair: Cristopher Camacho

 

14:00

14:40

 

Chiyung Yam, Beijing Computational Science Research Center, China
Theoretical investigation of current-induced light emission in scanning tunneling microscopy molecular junctions

 

14:40

-

15:20

 

Qiang Cui, Boston University, Massachusetts, USA
Improvement of DFTB model for condensed phase simulations

 

15:20

-

16:00

 

Jolla Kullgren, University of Uppsala, Sweden
Physically constrained splines – a step towards transferable repulsive potentials for SCC-DFTB

 

17:20

 

 

 

Poster Mounting

 

17:30

21:00

 

Poster Session, Catering Buffet (House of Science)

 

Friday, October 11th 2019 (House of Science Bremen, Downtown)

 

Session:

ML for electronic and spectroscopic properties

 

 

 

 

 

Chair: Balint Aradi

 

08:30

09:10

 

Patrick Rinke, Aalto University, Helsinki, Finland
ARTIST: artificial intelligence for spectroscopy

 

09:10

09:50

 

Olexandr Isayev, University of North Carolina, Chapel Hill, USA
Accurate and transferable multitask prediction of chemical properties with an atoms-in-molecule neural network

 

09:50

10:30

 

Anders M. N. Niklasson, Los Alamos National Laboratory, New Mexico, USA
Graph-based linear scaling electronic structure theory

 

10:30

11:00

 

Coffee Break

 

11:00

11:40

 

Sergei Tretiak, Los Alamos National Laboratory, New Mexico, USA
Multiple cloning and polaritons in excited state non-adiabatic molecular dynamics

 

11:40

12:20

 

Gotthard Seifert, Technical University of Dresden, Germany
Bridging scales in materials simulations - quantum versus
classical simulations

 

12:20

12:25

 

Closing words: Thomas Frauenheim

 

12:25

 

 

 

Departure