Latest update: 17 Dec 2018

Winter School in Theoretical Chemistry 2018

This is the 34^th Winter School

Machine Learning in Theoretical Chemistry

Helsinki, December 10^th–13^th, 2018

Lecturers

Jörg Behler, Georg-August-Universität Göttingen, Göttingen, Germany

Michele Ceriotti, EPFL, Lausanne, Switzerland

Volker Deringer, University of Cambridge, Cambridge, UK

Jason Goodpaster, University of Minnesota, Twin Cities, USA

Olexander Isayev, University of North Carolina, Chapel Hill, USA

Anatole von Lilienfeld, University of Basel, Basel, Switzerland

Paul Popelier, The University of Manchester, Manchester, UK

Milica Todorovic, Aalto University, Espoo, Finland

Contributing speakers

Volker Settels & Andrew Palmer, BASF SE, Germany

Michael Sluydts, University of Ghent, Belgium

Julia Westermayr, University of Vienna, Austria

Programme

Monday 10 December

9:00	Welcome
9:15	Behler 1	High-Dimensional Neural Networks: Concepts and Applications, Part 1
10:00	Behler 2	High-Dimensional Neural Networks: Concepts and Applications, Part 2
10:45	Break
11:00	Behler 3	High-Dimensional Neural Networks: Concepts and Applications, Part 3
11:45	Lunch
13:00	Sillanpää	HPC-Europa3: Travel and collaborate with EC funding
13:10	Sluydts	Accelerating materials screening with machine learning
13:30	Popelier 1	Next generation force field design: state of the art and challenges
14:15	Popelier 2	Background to FFLUX: Quantum Chemical Topology and Gaussian Processes
15:00	Break
15:15	Deringer 1	Machine-learning potentials for materials chemistry: fundamentals
16:00	Deringer 2	Machine-learning potentials for materials chemistry: applications to amorphous materials
16:45	Break
17:00	Poster Session

Tuesday 11 December

9:15	Goodpaster 1	Kernel Ridge Regression, Gaussian Processes, and Neural Networks in Quantum Chemistry
10:00	Ceriotti 1	Atom-density based representations for machine learning
10:45	Break
11:00	Ceriotti 2	Not only potentials: learning vectors and tensors
11:45	Lunch
13:00	Westermayr	Machine learning for excited-state molecular dynamics
13:20	Settels & Palmer	High-throughput forecasting of molecular properties in solution
13:45	Isayev 1	Predicting properties of inorganic materials with machine learning
14:30	Isayev 2	Neural Networks learning Quantum Chemistry
15:15	Break
15:30	Popelier 3	Results and Future Work
16:15	Todorovic 1	Global atomistic structure search with Bayesian Optimization

Wednesday 12 December

9:15	Lilienfeld 1	Quantum Machine Learning in Chemical Space: Part 1
10:00	Lilienfeld 2	Quantum Machine Learning in Chemical Space: Part 2
10:45	Break
11:00	Goodpaster 2	Learning Electron Correlation: Part 1
11:45	Goodpaster 3	Learning Electron Correlation: Part 2
12:30	Lunch and free afternoon
13:00	(Annual meeting of the Computational Chemistry Section of the Finnish Chemical Societies)

Thursday 13 December

9:15	Todorovic 2	Predicting molecular orbital energies with Kernel Ridge Regression
10:00	Todorovic 3	Deep-learning molecular spectra with Neural Networks
10:45	Break
11:00	Deringer 3	Data-driven learning and prediction of inorganic crystal structures
11:45	Isayev 3	Deep Learning and Generative Models for Inverse Molecular Design
12:30	Closing and departure

Further Information

Place:

The lectures will take place in Lecture Hall A110, Department of Chemistry, Kumpula Campus, University of Helsinki. Google Maps might be helpful.

Registration:

Registration is closed. Registration closed at midnight local time, 30 November. Note: Some e-mail addresses have been entered incorrectly during the registration. If you have registered, but did not receive, say, the welcome e-mail sent December 4^th, then there are some problems with your mail address.

Posters and abstracts:

A poster session will be held on Monday evening. The posters are browsable throughout the duration of the School.

Abstracts:

Dead-line for submitting abstracts has passed. Check that your poster abstract is correctly included in the Abstract book (preliminary).

Lecture notes:

Notes are stored behind a password here: Lecture Notes. More might be added within time.

Credits:

Students who wish to have credits for the course, should write an essay on one of the School topics.

The essay titles are:

Machine learning potentials

Machine learning and electron correlation

Predicting properties of materials by using machine learning

Predicting molecular properties by using machine learning

Machine learning in chemical space

Models for inverse design of molecules

Design strategies for force fields

High-dimensional neural networks and optimization strategies

Predicting crystal structures using machine learning

Molecular dynamics simulations and machine learning

Your own title related to the theme of the winter school

The dead-line for submitting the essay is January 31^st, 2019.

Send a PDF of your essay to winterschool@chem.helsinki.fi. The course will give 3 ECTS.

Fees:

The School is free of charge for academic staff and students

Accommodation:

Participants should make their own accommodation arrangements.

List of Hotels in Helsinki

Questions:

General questions concerning the School can be addressed to Dage Sundholm or Mikael Johansson

Address:	Department of Chemistry P.O. Box 55 (A.I. Virtasen aukio 1) FI-00014 University of Helsinki Finland
Email:	winterschool@chem.helsinki.fi

How to find the campus:

See Chemicum on Google Maps

Practical notes:

Some things to consider.

Arriving at the airport, you can take the Finnair bus (6.90 €, for customers of all airlines) or the new train connection between the airport and the central railway station with ringrail trains I and P (5.50 €).
The Journey Planner for public transport, and Google Maps might be helpful. Wikipedia knows about public transport in Helsinki as well.
A taxi from the airport to the Kumpula campus (A.I. Virtasen Aukio 1) or the city centre should be around 40–50 €. Note: As of July 2018, taxi prices are unregulated, so a ride can in principle cost anything. Best to ask the driver beforehand! This applies to all taxi rides, not only from/to the airport.
Look at the latest weather reports before travelling as temperatures might drop down to -15 degrees for the Winter School week. This year, the temperatures seem to be quite evenly zero °C.

Eat and drink:

Lunch is available in the various canteens of the campus. For the evenings, we have collected a list of places suitable for slightly larger groups. The ones in italics are also good for just a drink or three.

Other things of interest:

Helsinki might be cold, but is not dead in December. Check out the #myhelsinki pages.