Stochastic Approximation

SS 2021




Tue. 10 a.m. to 12 a.m.

Fr. 10 a.m. to 12 a.m.

Dozent:  Prof. Dr. Steffen Dereich
Assistenz:  Sebastian Kassing

This lecture in the course overview

These tutorials in the course overview

Inhalt: The lecture "Stochastic Approximation" is aimed at Master's students who have a sound basic knowledge of probability theory including martingale theory. The lecture is concerned with stochastic approximation. This involves the analysis of stochastic processes in discrete time, whose behaviour is closely linked to a deterministic ordinary differential equation. In applications, such stochastic processes occur, for example, in iterative, stochastic optimisation algorithms.

The lecture introduces techniques for the asymptotic analysis of stochastic processes and then applies these to analyse special processes from the field of stochastic approximation. In particular, the Robbins-Monro algorithm and its Polyak-Ruppert smoothing will be considered.
Stochastic approximation is a technique that is frequently used (also in recent publications), and the topics of the lecture form a good basis for the assignment of a master's thesis topic.

Learnweb: The corresponding learnweb course is here to be found.

Tutorials: We. 12 p.m. to 14 p.m.