Bjoern Schaefer
Zentrum für Astronomie, Universität Heidelberg (Heidelberg, Allemagne)
Amphithéâtre Henri Mineur, Institut d'Astrophysique de Paris
|
| « Bayesian inference in cosmology with partition functions » |
|
Bjoern Schaefer |
| Bayesian inference in cosmology often deals with posterior parameter distribution of complicated shape and with quantities that are difficult to estimate from a sampling process. In my talk I hope to show that, using parallels and analogies between statistical mechanics and information theory, partition functions are analytic descriptions of sampling processes which allow the computation of many quantities of interest, with suitable numerical methods, borrowing ideas from AI. I'll end in an outlook to Bayesian model selection and what partition functions could achieve in this application. |
|
vendredi 8 novembre 2024 - 11:00 Amphithéâtre Henri Mineur, Institut d'Astrophysique de Paris |
| Page web du séminaire / Seminar's webpage |