Symposium »Concepts of Probability in the Sciences« Mitschnitt einer Veranstaltung des Erwin Schrödinger International Institute for Mathematics and Physics (ESI) am Dienstag, dem 30. Oktober 2018 im Boltzmann-Hörsaal des ESI Teil 14. Gerhard Hummer: "Inferring statistical ensembles from experiment and simulation" Kamera: Matthias Dellago Schnitt: Daniel Winkler Gerhard Hummer ist Professor am Max-Planck-Institut für Biophysik in Frankfurt am Main. Abstract: Statistical mechanics describes the thermal disorder of molecular systems in terms of statistical ensembles, such as the Boltzmann distribution of classical mechanics. Given an approximate classical or quantum mechanical energy function, molecular simulations produce weighted configurations that jointly represent the statistical ensemble. How should one adapt the weights of these configurations, or more generally the underlying “Boltzmann distribution”, when new experimental information emerges? I address this problem of “ensemble refinement” in a (somewhat unusual) Bayesian framework. INHALT ====== Kapitel Titel Position --------------------------------------------------------------------- 1. Vorspann 00:00:00 2. Objective: integrative structure biology 00:00:08 3. Ensemble theory. Bayesian inference 00:06:52 4. Likelihood functions 00:15:50 5. Ensemble refinement as a "mechanical" problem 00:23:11 6. Spin-label rotamer states in resonance spectroscopy 00:32:06

Statistische Physik, Ensemble, Bayes-Statistik

video/mp4

MP4-kompatibler Mediaplayer

PT38M15S

(c) Österreichische Zentralbibliothek für Physik