Universität Wien

450102 SE COSMOS Topic Seminar: Bayesian Approaches to Inverse Problems in Astrophysics and Cosmology (2024W)

3.00 ECTS (2.00 SWS), SPL 45 - Doktoratsstudium Geowissenschaften, Geographie und Astronomie
Prüfungsimmanente Lehrveranstaltung
Moodle
Mo 09.12. 13:15-14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17

An/Abmeldung

Hinweis: Ihr Anmeldezeitpunkt innerhalb der Frist hat keine Auswirkungen auf die Platzvergabe (kein "first come, first served").

Details

max. 10 Teilnehmer*innen
Sprache: Englisch

Lehrende

Termine (iCal) - nächster Termin ist mit N markiert

  • Montag 07.10. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 14.10. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 21.10. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 28.10. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 04.11. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 11.11. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 18.11. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 25.11. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 02.12. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 16.12. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 13.01. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 20.01. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17
  • Montag 27.01. 13:15 - 14:45 Seminarraum 2 Astronomie Sternwarte, Türkenschanzstraße 17

Information

Ziele, Inhalte und Methode der Lehrveranstaltung

Bayesian hierarchical modelling is a powerful and flexible technique for data modelling problems. In this framework, we build probabilistic models of the data generating process (e.g. including physical effects, selection functions, observational noise etc) and use this model to infer unobserved model parameters and to predict unseen data. This approach provides a theoretically principled way to quantify uncertainty, incorporate prior information, and combine datasets.

The aim of this course is to equip students with knowledge and computational tools to apply Bayesian methods to problems in astrophysics and cosmology. By the end of the course, students should be able to (i) explain the concepts behind Bayesian, hierarchical modelling, (ii) devise bespoke probabilistic models for data modelling problems, and (iii) execute a range of methods of statistical inference.

The course will start with five lectures which introduce students to the theoretical framework and computational tools (specifically, the probabilistic programming language numpyro). Students will then apply this knowledge by splitting into pairs to work on hands-on projects. A number of projects spanning problems in astrophysics and cosmology will be provided by the lecturers. Throughout the course, lecturers will be available to discuss progress and offer advice.

Art der Leistungskontrolle und erlaubte Hilfsmittel

Assessment will be out of 100 points:
- 30 points for a final presentation of the project (with slides),
- 60 points for a written project report (7-8 pages),
- 10 points for active participation in discussions and the lectures.
Detailed criteria for evaluating the presentations and written report will be provided in the course.

Mindestanforderungen und Beurteilungsmaßstab

Students must submit a final presentation and project report, and they must attend all lectures and final presentations (any difficulties should be communicated in advance). Attendance during the period of project-work is not mandatory.

Prüfungsstoff

n/a

Literatur

Resources will be discussed in the course.

Zuordnung im Vorlesungsverzeichnis

BEN; VAF; PM-Astr; PM-FnNawi;

Letzte Änderung: Mo 07.10.2024 11:07