Topic: 
Intelligente Energieversorgungsnetze 
Date: 
08.11.21 
Time: 
16:15 
Place: 
H4 
Guest: 

FH Bielefeld 

Abstract: 

Contact person: 
Topic: 
Random matrices, spin glasses, and machine learning 
Date: 
23.07.21 
Time: 
16:15 
Place: 
ZOOM/Konferenzschaltung 
Guest: 

Oxford University 

Abstract: 
I will describe some problems relating to machine learning and their connections to random matrix theory and spin glasses. These connections give a mathematical framework for understanding in qualitative terms the effectiveness of certain algorithms that are important in machine learning, but developing them into precise models remains a major challenge. I will reflect on the different roles played by models in computer science and physics, focussing on those involving random matrices. 
Contact person: 
Topic: 
Machine Learning for Thermodynamic Observables in Lattice Field Theories 
Date: 
06.07.21 
Time: 
14:15 
Place: 
Online, via ZOOM 
Guest: 

Perimeter Institute, Ontario, Canada 

Abstract: 
In this talk, I will discuss how applying machine learning techniques to lattice field theory is a promising route for solving problems where Markov Chain Monte Carlo (MCMC) methods are problematic. More specifically, I will show that deep generative models can be used to estimate thermodynamic observables like the free energy, which contrasts with existing MCMCbased methods that are limited to only estimate free energy differences. I will demonstrate the effectiveness of the proposed method for twodimensional $\phi^4$ theory and compare it to MCMCbased methods in detailed numerical experiments. 
Contact person: 
Topic: 
14:30 Untersuchung von frustrierten Spin1/2Systemen mit Hilfe von quantumthreecoloring am Beispiel des Kuboktaeders 
Date: 
14.10.21 
Time: 
14:30 
Place: 
Hybrid  Zoom/D5153 
Guest: 
Florian Brökemeier 
Universität Bielefeld 

Abstract: 

Contact person: 
Topic: 
On NonHermitian BetaEnsembles 
Date: 
14.10.21 
Time: 
16:00 
Place: 
D5153 
Guest: 

Universität Bielefeld 

Abstract: 
Loggases with inverse temperature beta are systems with many applications in physics, for example in the theory of superconductors or the fractional quantum Hall effect. For some specific values of beta a correspondence to random matrix theory (RMT) is well established. The advantage of this connection is the usage of the RMT methods in the study of those systems. The goal of this talk is the discussion of Loggases in two dimensions, i.e. in the nonHermitian case, for more general values of the inverse temperature. Therefore, we study in the first part a model of normal 2 × 2 matrices with beta in [0,2] and discuss whether we find a surmise for the nearestneighbour spacing distribution of large matrices. In the second part of the talk we introduce the study of symmetry classes in nonHermitian RMT. We conjecture that the classes of complex symmetric and complex quaternion matrices can be effectively described by Loggases in two dimensions with noninteger inverse temperatures. 
Contact person: 
Topic: 
Planar symplectic ensembles: from scaling limits to Wronskian structures 
Date: 
27.10.21 
Time: 
09:00 
Place: 
ZOOM / Konferenzschaltung 
Guest: 

Korea Institute for Advanced Study 

Abstract: 
In this talk, I will discuss complex eigenvalues of nonHermitian random matrices with symplectic symmetry, which are known to form Pfaffian point processes. In particular, I will present various scaling limits of symplectic ensembles and explain their unified integrable structure of Wronskian form. Examples include edge scaling limits of the Ginibre ensemble (with boundary confinements) and bulk/edge scaling limits of the elliptic Ginibre ensemble in the almostHermitian regime. Beyond standard universality classes, I will also introduce scaling limits of the MittagLeffler ensemble at the singularity. Furthermore, for symplectic ensembles with general external potentials, I will present the characterization of translation invariant scaling limits by virtue of rescaled massone and Ward’s equations. This talk is based on two joint works: one with Gernot Akemann and NamGyu Kang and the other with Markus Ebke and SeongMi Seo. 
Contact person: 