Weitere Fächer
Wissenschaftliche Direktorin: Prof. Dr. Herta Flor
Tel.: 0621 1703-6302, E-Mail
Sekretariat: Angelika Bauder
Tel.: 0621 1703-6302, E-Mail
Institut für Neuropsychologie und Psychologie
Abteilung Klinische Psychologie
Leitung: Prof. Dr. Peter Kirsch
Tel.: 0621 1703-6501, -6511, E-Mail
Sekretariat: Ellen Schmucker
Tel.: 0621 1703-6502, E-Mail
Abteilung Klinische Psychologie
Lehrende / Instructors:
- apl. Prof. Dr. Josef Bailer - LSF
- Prof. Dr. Herta Flor - LSF / Uni Mannheim
- Prof. Dr. Peter Kirsch - LSF
- Dr. Patric Meyer, Dipl.-Psych. - LSF / Uni Mannheim
- Dr. Frauke Nees, Dipl.-Psych. - LSF / Uni Mannheim
- Dr. Jörg Trojan, Dipl.-Psych.
Wissenschaftlicher Direktor: Prof. Dr. Rainer Spanagel
Tel.: 0621 1703-6251, E-Mail
Sekretariat: Christine Roggenkamp
Tel.: 0621 1703-6252, E-Mail
Institut für Psychopharmakologie
Lehrende:
Veranstaltungshinweise:
Aktuelle Informationen zum Seminar für Psychopharmakologie finden Sie auf der Seite Veranstaltungen. / For up-to-date information about the Psychopharmacology Seminar, please visit the Events page.
Professor für Theoretische Neurowissenschaften
Abteilungsleiter: Prof. Dr. Daniel Durstewitz
Tel.: 0621 1703-2361, E-Mail
Sekretariat: Christine Roggenkamp, M.A.
Tel.: 0621 1703-6252, E-Mail
Abteilung Theoretische Neurowissenschaften
Lehrende:
Veranstaltungen im Wintersemester 2022/2023
MVSpec Lecture Dynamical Systems Theory in Machine Learning & Data Science
- Time: Wed 11.00‐13.00 (lecture), Wed 14.00 – 16.00 (exercises)
Location: INF 227, SR 2.403 - Lecturers: D. Durstewitz, Z. Monfared
Dynamical systems theory (DST) provides a powerful and beautiful mathematical framework for understanding the behavior of natural or engineered systems that evolve in time (and potentially space), usually modeled by sets of differential equations or time-recursive maps. Such systems produce a range of universal and common phenomena, such as different types of attractor states, routes to chaos, synchronization and nonlinear oscillations, or bifurcations. Bifurcations are particularly important to understand phenomena like Covid19 disease propagation or climate change
In recent years, DST has become a hot topic in machine learning and AI research. For one, it offers mathematical tools for understanding and improving training processes in deep learning. On the other hand, recurrent neural networks (RNNs), as commonly used to model and predict sequential and time series data, are nonlinear dynamical systems themselves whose behavior can be analyzed by DST tools.
RNNs may also be used to directly learn the underlying dynamical system from time series observations. Finally, we will discuss recent developments like “neural ordinary differential equations” that provide powerful approximation frameworks based on DST ideas.
Veranstaltungen im Sommersemester 2022
MVSpec lecture ‘Time Series Analysis & Recurrent Neural Networks’
- Time: Wed 11.00‐13.00 (lecture), Wed 14.00 – 16.00 (exercises)
Location: INF 227, SR 2.403 - Lecturers: D. Durstewitz, N.N.
This course will deal primarily with model‐based analysis of time series, that is with insights and predictions that could be gained by inferring a mathematical model of the dynamical process from the observed data.It will cover state of the art methods from the fields of computational statistics, machine learning & AI, and nonlinear dynamics.
Starting from simple linear auto‐regressive models, we will advance to nonlineardynamical systems, state space approaches, and generative deep recurrent neural networks. The latter class of models is particularly interesting and powerful, as it can – after being trained on time series data – generate new instances of the observed system’s behavior on its own, e.g. new samples of text written in a certain style, or new trajectories from an observed dynamical system.
Zentralinstitut für Seelische Gesundheit (ZI) - https://www.zi-mannheim.de