Welcome to the Computer Tutorial in Markov Modeling WS 2020/21

emma2015_sm.png deeptime logo

News

23 Feb 2021: The course is booked out.

05 Feb 2021: This year's workshop will for the first time feature PyEMMA and DeepTime.

Dates

5-day course 01.-05.03.2021 9:00-15:00 online

Program

In this 1-week compact course (Blockseminar) participants will learn how to quantitatively analyze time-series data, with a special focus on molecular dynamics simulations. Due to the COVID-19 pandemic, the course will take place online and will include both theory lessons (20-30%) as well as practical applications using the software packages PyEMMA and DeepTime. All computer applications will be done in Python and Jupyter notebooks. We will provide test data, but participants may also bring their own data (any standard format, e.g., xtc, dcd) and we will help with applying PyEMMA or DeepTime to it.

Schedule

Overview

Theory lessons and associated software application sessions take place Monday to Wednesday. Thursday morning is dedicated to contributed talks. There is an open space for project work from Thursday afternoon to Friday noon. During this time, we will be available to support external scientists with their own projects and FU students with project tasks (to be presented on Friday afternoon).

Schedule

The time is in local Berlin time (GMT+1).

  Mon Tue Wed Thu Fr
09-12 lectures/tutorials lectures/tutorials lectures/tutorials contributed talks projects
12-13 lunch break lunch break lunch break lunch break lunch break
13-15 lectures/tutorials lectures/tutorials lectures/tutorials projects* student talks


*On Thursday afternoon, assistants will be available until 17:00 Berlin time to assist with projects.



Sessions

Monday

Welcome (5 min, Tim Hempel)

MSM theory (50 min, Tim Hempel)

PyEMMA /DeepTime package development / intro to installation session (10 min, Moritz Hoffmann)

Installation (45 min)

[Break 15 min]

data I/O, featurization, dimension reduction, clustering incl. tutorial (60 min, Moritz Hoffmann)

[lunch break ~1h]

MSM estimation + validation tutorial (60 min, Lluís Raich)

Welcome address (45 min, Frank Noé)

Tuesday

Brief repetition (10 min, Tim Hempel)

VAMP theory (20-30 min, Moritz Hoffmann)

I/O, Vamp, MSM estimation mini project (30+ min)

MSM analysis + CG + TPT incl. tutorial (120 min, Luís Raich)

[lunch break ~1h]

MSM analysis mini project (30 min)

HMM theory and tutorial (60 min, Tim Hempel)

Wednesday

Brief repetition (10 min)

SINDy and other Koopman-based methods (EDMD, kEDMD, kCCA) (2h 50min, Moritz Hoffmann)

[lunch break ~1h]

VAMPNets intro (30 min, Andreas Mardt)

VAMPNets mini projects (60 min)

Lectures wrap-up (10 min)

Preparation for course

Installation of PyEmma

For tutorials, we will use the Python-based software packages PyEMMA (Emma's Markov Model Algorithms, http://pubs.acs.org/doi/abs/10.1021/acs.jctc.5b00743, http://github.com/markovmodel/PyEMMA) and DeepTime (https://github.com/deeptime-ml/deeptime). We strongly recommend to install both packages from the anaconda Python distribution via conda install -c conda-forge pyemma deeptime
If you are not using the anaconda Python distribution, PyEMMA and DeepTime are also available via source installation from Github. For more detailed installation instructions, please refer to http://www.emma-project.org/latest/INSTALL.html and .
If possible, install the latest release of PyEMMA.

Information for students

General

Course ID: 20112011 Language: English
SWS:2; LP/Credits: 5

Registration

FU students who were unable to register in Campus Management but registered with us should do a manual "Modulanmeldung" after the course. You find the form on the Registrar's office website. Complete and drop into mail box outside of registrar's office. Once you've been added to the course roster we can confirm your successful participation.

Credit Requirements (Scheinkriterien)

  • Active participation: Students will get project assignments (usually teams of two) on Wednesday, will work on their project on Thursday afternoon and Friday morning, and present on Friday afternoon (15 mins per team, both students should present).
  • For physics students, this course counts as a seminar "Selected Topics in Physics" (5 LP)
  • All participants will receive a certificate of attendance after the workshop by email.
  • FU students will get their credits assigned in Campus Management
  • Students from other unis can pick up a paper certificate (Schein) stating their credits on the last day of the workshop.

Information for non-resident applicants

As every year, we can offer a limited number of placements for PhD students and PostDocs interested in applying PyEmma or DeepTime to their simulations. To inquire about free spots, send an email to moritz.hoffmann[at]fu-berlin.de stating your affiliation and research focus. There is no official registration required. We will put you on the course roster and mailing list.



Information for participating

We will post the links to participate here.

Contact

Name   Mail
Moritz Hoffmann coordinator moritz.hoffmann[at]fu-berlin.de
Tim Hempel coordinator tim.hempel[at]fu-berlin.de

This topic: CompMolBio > MarkovSeminar21
Topic revision: 26 Feb 2021, MoritzHoffmann