Workshop on state of the art in sampling and clustering

5 Oct 2020, 12:00 → 9 Oct 2020, 17:00 Europe/Berlin

Zoom, Gather Town and MPP Auditorium (Virtually and at MPP)

Zoom connection information and other details will be sent to participants via Indico reminder emails each morning. Watch out for emails with the subject "[Indico] [Event reminder] Workshop on state of the art in sampling and clustering", check your spam folder as well if necessary. Please contact us if you didn't receive a first such email yet.

The workshop will cover both introductory and advanced topics in the field of statistical sampling and clustering. In addition to lectures on state of the art approaches, the workshop will also comprise hand-on and excercise sessions.

The workshop, hosted by the Max Planck Institute for Physics (MPP)  is organised by the INSIGHTS ITN, MPP IMPRS and the ORIGINS Excellence Cluster, it is open to everybody affiliated with these organizations.

Due to Covid-19, all sessions and social interactions will be held online. Some of the speakers and a limited number of people will participate in person at the MPP auditorium.

Note: Registration for this event is now closed.

Monday, 5 October

13:00 → 13:10 Welcome

Speakers: Allen Caldwell (Max Planck Institute for Physics), Oliver Schulz (Max Planck for Physics)

13:10 → 15:00 Introduction to Information Field Theory (IFT)

Speaker: Torsten Enßlin (Max Planck Institute for Astrophysics)

IFT-2020-10-05.pdf

15:00 → 15:30 Virtual Social Interaction/Discussion

15:30 → 17:00 NIFTy – Numerical Information Field Theory

NIFTy "Numerical Information Field Theory", is a versatile library designed to enable the development of signal inference algorithms that are independent of the underlying grids (spatial, spectral, temporal, …) and their resolutions. Its object-oriented framework is written in Python, although it accesses libraries written in C++ and C for efficiency.

Speaker: Philipp Arras

NIFTy6 introduction

Tuesday, 6 October

09:00 → 10:30 Approximate Bayesian Computation (ABC)

Speaker: Christian Robert (Université Paris Dauphine PSL)

ABCourse.pdf

10:30 → 12:00 Approximate Bayesian Computation (ABC)

Speaker: Christian Robert (Université Paris Dauphine PSL)

12:00 → 13:30 Break

13:30 → 14:15 Statistics in autonomous driving

Speaker: Frederik Beaujean (MPP)

14:15 → 15:00 Virtual Social Interaction/Discussion

15:00 → 16:00 Epidemic Models to Quantify the Effects of Testing, Contact Tracing and Containment

Speaker: Manuel Gomez Rodriguez (MPI for Software Systems)

Wednesday, 7 October

10:00 → 11:30 Foundations of Clustering

Speaker: Debarghya Ghoshdastidar (TUM)

MPP-Clustering-Debarghya.pdf

11:30 → 12:00 Virtual Social Interaction/Discussion

12:00 → 13:30 Break

13:30 → 15:00 Clustering, continued

Speaker: Debarghya Ghoshdastidar (TUM)

15:00 → 15:30 Virtual Social Interaction/Discussion

15:30 → 17:00 Introduction to nested sampling

Speaker: Johannes Buchner (MPI for Extraterrestrial Physics)

2020-nested_sampling.pdf

Tutorial 1

Tutorial 2

Tutorial 3

Tutorial 4

Tutorial 5

Thursday, 8 October

10:00 → 12:00 Clustering, hands-on

Speaker: Philipp Eller (Max Planck for Physics)

Clustering_II.ipynb

Clustering_I.ipynb

12:00 → 13:30 Break

13:30 → 15:00 Introduction to Hamiltonian Monte Carlo (HMC)

Speaker: Michael Betancourt

Markov Chain Monte Carlo in Practice

markov_chain_monte_carlo.pdf

Probabilistic Computation

15:00 → 15:30 Virtual Social Interaction/Discussion

15:30 → 17:00 More on Hamiltonian Monte Carlo, with exercises

Speaker: Michael Betancourt

Friday, 9 October

10:00 → 11:00 Updates on BAT.jl, a Bayesian Analysis Toolkit in Julia

BAT.jl is a Bayesian Analysis Toolkit implemented in the Julia language. It is a high high-performance tool box for Bayesian inference with statistical models expressed in a general-purpose programming language, instead of a domain-specific language.

Typical applications for this package are the extraction of the values of the parameters of a model, the comparison of different models in the light of a given data set and the test of the validity of a model to represent the data set at hand. BAT.jl provides access to the full Bayesian posterior distribution to enable parameter estimation, limit setting and uncertainty propagation. BAT.jl also provides supporting functionality like plotting recipes and reporting functions.

Speaker: Oliver Schulz (Max Planck for Physics)

BAT Documentation

BAT.jl environment

oschulz-bat.pdf

11:00 → 12:00 Social Discussion Session