Institutionen för naturvetenskap och teknik

Research Seminars in Mathematics

The research seminars are in the subjects of pure, applied, and computational mathematics and are usually held at the afternoons on Fridays. All are welcome to attend!

Please contact Jakob Palmkvist if you have any questions regarding this seminar series.


Speaker: Sweta Das, Örebro University

Date: Friday 12th April, 13:15

Location: T213

Title: "Minimal Degenerations of Eigenstructure for Skew-Symmetric Matrix Polynomials"

Abstract: In this presentation, we describe qualitatively how eigenstructure of skew-symmetric matrix polynomials of odd degree changes under small perturbations in the matrix coefficients. Using strong linearization we prove a necessary and sufficient condition for one orbit of the linearization of a matrix polynomial to be a proper subset of the closure of the orbit of linearization of another polynomial. To achieve this, we introduce a set of rules describing structure transitions of the canonical blocks of the polynomial's linearization. These rules facilitate the construction of the stratification graphs of linearization of the polynomials. Finally, we state a method that allows us to sketch the entire or a part of the stratification graph of one matrix polynomial's linearization in relation to another matrix polynomial's linearization, provided both polynomials share the same degree but have different dimensions. 

Speaker: Hugo Strand, Örebro University

Date: Friday, December 8 at 13.15, T211, Teknikhuset

Title: Quasi Monte Carlo for open quantum many-body systems

Abstract: The simulation of open quantum many-body systems is a formidable challenge. Due to the dimensionality of the problem the goto method is stochastic Markov chain Monte Carlo, which converges as N^{-1/2} where N is the number of samples. We show how Markov chain Monte Carlo can be replaced by quasi Monte Carlo integration in the inchworm method, improving the convergence rate to N^{-1} (1). This extends the applicability of the inchworm method to materials simulation, where continuous time hybridization expansion Monte Carlo has a severe sign problem. A Julia implementation of our inchworm quasi Monte Carlo approach: "QInchworm.jl", is available under an open source license (2).

1) Inchworm quasi Monte Carlo for quantum impurities

2) QInchworm

Speaker: Volker Mehrmann

Date: Wednesday, December 6 at 13.15, Bio in Forumhuset

Title: Energy Based Mathematical Modeling, Simulation, and Control of Real World Systems

Abstract: Most real world dynamical systems consist of subsystems from different physical domains, modelled by partial-differential equations, ordinary differential equations, and algebraic equations, combined with input and output connections. To deal with such complex system, in recent years the class of dissipative port-Hamiltonian (pH) descriptor systems has emerged as a very successful modeling methodology. The main reasons are that the network based interconnection of pH systems is again pH, Galerkin projection in PDE discretization and model reduction preserve the pH structure and the physical properties are encoded in the geometric properties of the flow as well as the algebraic properties of the equations. Furthermore, dissipative pH system form a very robust representation under structured perturbations and directly indicate Lyapunov functions for stability analysis.

Another advantage of energy based modelling via pH systems is that each separate model of a physical system can be a whole model catalog from which models can be chosen in an adaptive way within simulation and optimization methods.

We discuss the class of constrained pH systems and illustrate how many classical real world mathematical models can be formulated in this class.

We illustrate the results with some real world examples from gas transport and district heating systems and point out emerging mathematical challenges.

Speaker: Anna Oleynik, University of Bergen

Date: Friday, November 3 at 13.15, T209, Teknikhuset

Title: Applied mathematics for marine monitoring

Abstract: I will talk about how simple and complex models can assist in marine pollution monitoring and fish stock assessment. Specifically, I will focus on the tracking of plastics in Norwegian fjords, employing the Lagrangian particle tracking model. Next, I will introduce a novel mixed-integer programming problem designed for optimizing path planning, and I will showcase its practical application in the context of marine carbon storage monitoring. Additionally, I will present a Deep Learning approach for the classification of acoustic signals from marine surveys. At the end, I will dedicate some time to the projects in collaboration with Örebro University.

Speakers: Magnus Ögren, Daniel Edström

Date: Friday, September 15 at 13.15, T207, Teknikhuset

Title: Some random calculations for some non-randomly chosen PDE problems, with applications within porous media technologies

Abstract: Vi diskuterar några exempel på problem med olika randvillkor där vi söker stationära och tidsberoende lösningar, och gradienter till lösningar, av Laplaceekvation i ’krångliga’ geometrier. Genomgången baseras på personliga erfarenheter med kompling till teknologi, men för den som vill titta på den teoretiska bakgrunden är Feynman-Kacs formel och teorin kring Fokker-Planck ekvationer nyttig. Metoder med slumpvandringar (vilka kan ses som diskreta implementationer av SDEs) kan ha fördelar i: högre dimensioner och stora domäner, där ’filterfel’ av data kan göra den ursprungliga domänen felaktig, komplicerade (rörliga) randvillkor, eller där analysens språkbruk inte på enklaste sätt modellerar verkligheten. Exempel av ovanstående kommer att diskuteras.

Speakers: Daniel Blixt, Scuola Superiore Meridionale, Italy.

Date: Monday, May 22 at 10.15, T211, Teknikhuset.

Title: Teleparallel gravity

Abstract: Since Einstein formulated his theory of general relativity it has transitioned from a controversial theory to a widely accepted theory supported by observations. At early times it was confirmed by solar system tests, and recently we have made the first observations of gravitational waves and the first image of a black hole, both of them predicted by general relativity. A very famous interpretation of the equations governing general relativity reads "Spacetime tells matter how to move; matter tells spacetime how to curve''. Is this the only way to think about it? In this talk I will introduce teleparallel theories of gravity, where torsion plays the central role, and it turns out that general relativity can equivalently be formulated in this framework. In recent years there has been an increased interest in modified theories of gravity. We live in an interesting era of astronomy, where it is now much easier to compare predictions of general relativity and modified gravity. Furthermore, we are finding growing tensions in cosmology. Could this be related to an incomplete understanding of gravity? Could teleparallel gravity provide us with a better theory of gravity?

Speakers: Jasur Yuldashev, Uzbek Academy of Sciences, Tashkent, Uzbekistan.

Date: Friday, May 12 at 13.15, T207, Teknikhuset.

Title: Management of solitons in medium with competing cubic and quadratic nonlinearities

Abstract: Management of solitons in media with competing quadratic and cubic nonlinearities is investigated. Two schemes, using rapid modulations of a mismatch parameter, and of the Kerr nonlinearity parameter are studied. For both cases, the averaged in time wave equations are derived. In the case of mismatch management, the region of the parameters where stabilization is possible is found. In the case of Kerr nonlinearity management, it is shown that the effective χ(2) nonlinearity depends on the intensity imbalance between fundamental (FH) and second (SH) harmonics. Predictions obtained from the averaged equations are confirmed by numerical simulations of the full PDE's.

Speakers: Pavel Bessarab, Linnaeus University, Kalmar.

Date: Friday, April 14, Zoom meeting.

Title: Identification of transition paths for the prediction of thermal stability and optimal control of magnetic systems

Abstract: The identification of paths that are optimal in some sense is an important problem in many areas of science. In this talk, the focus will be on paths for magnetic transitions. An important example is minimum energy paths (MEPs), paths in configuration space that connect states with different orientation of magnetic moments and lie low on the energy surface. MEPs have maximal statistical weight and, therefore, represent mechanisms of thermally activated transitions. Combined with the development of harmonic transition state theory, MEP calculations make it possible to predict the lifetime of magnetic states at a given temperature. MEPs do not, however, reflect the dynamics and there are other paths that represent minimal energy cost of switching between magnetic states. Such 'optimal control paths' (OCPs) involve rotation of magnetic moments in such a way that the system's internal dynamics aids the magnetic transition. The development of theoretical tools for finding MEPs and OCPs on high-dimensional energy surfaces will be discussed and applications presented where thermal stability of magnetic skyrmions is analyzed and spin-wave assisted magnetization switching in nanowires revealed.

Speakers: Finn Rietz and Jean-Paul Ivan, AASS, Örebro University.

Date: Friday, February 10 at 13.15, T215, Teknikhuset.

Title: Towards Prioritized Policy Composition + Gaussian Processes in Machine Learning: Constrained Regression and Inverse Problems

Abstract: (Finn Rietz) In Reinforcement Learning, we want to learn behavior policies that solve a sequential problem of interest by maximization of the corresponding reward function. More specifically, we investigate how we can compose multiple such behavior policies, which is not straightforward in the general case, although highly desirable: Composable RL policies allow for knowledge transfer, data reuse, modular design, and zero-shot or few-shot adaption to more and more complex tasks. In this seminar talk, I will introduce our approach for prioritized policy composition, what makes it challenging, and show some preliminary results.

Abstract: (Jean-Paul Ivan) A perspective on Gaussian processes that is particularly useful in machine learning is the view that a Gaussian process specifies a Bayesian prior over a function space. The posterior distribution resulting from conditioning on observations of any linear functional on this space is available in closed form, leading to an interpretable and flexible approach to nonlinear regression problems. This talk will present this approach to regression and show how this it allows ill-posed inverse problems to be reframed as well-posed inference problems.

Speaker: Massimiliano Fasi, Durham University, UK.

Date: Thursday, December 15, at 13.15, T213, Teknikhuset

Title: Computational Graphs for Matrix Functions

Abstract: Many numerical methods for evaluating matrix functions can be naturally viewed as computational graphs. Rephrasing these methods as directed acyclic graphs (DAGs) is a particularly effective approach to study existing techniques, improve them, and eventually derive new ones. The accuracy of these matrix techniques can be characterized by the accuracy of their scalar counterparts, thus designing algorithms for matrix functions can be regarded as a scalar-valued optimization problem. The derivatives needed during the optimization can be calculated automatically by exploiting the structure of the DAG, in a fashion analogous to backpropagation. GraphMatFun.jl is a Julia package that offers the means to generate and manipulate computational graphs, optimize their coefficients, and generate Julia, MATLAB, and C code to evaluate them efficiently at a matrix argument. The software also provides tools to estimate the accuracy of a graph-based algorithm and thus obtain numerically reliable methods. For the exponential, for example, using a particular form (degree-optimal) of polynomials produces implementations that in many cases are cheaper, in terms of computational cost, than the Padé-based techniques typically used in mathematical software. This is joint work with Elias Jarlebring (KTH) and Emil Ringh (Ericsson Research).

Speaker: Francesco Tudisco, GSSI Gran Sasso Science Institute, L’Aquila, Italy

Date: Friday, December 9, at 13.15, Zoom meeting

Title: Fast and efficient neural networks’ training via low-rank gradient flows

Abstract: Neural networks have achieved tremendous success in a large variety of applications. However, their memory footprint and computational demand can render them impractical in application settings with limited hardware or energy resources. At the same time, overparametrization seems to be necessary in order to overcome the highly nonconvex nature of the optimization problem. An optimal trade-off is then to be found in order to reduce networks’ dimensions while maintaining high performance.

Popular approaches in the literature are based on pruning techniques that look for “winning tickets”, smaller subnetworks achieving approximately the initial performance. However, these techniques are not able to reduce the memory footprint of the training phase and can be unstable with respect to the input weights. In this talk we will present a training algorithm that looks for “low-rank lottery tickets” by interpreting the training phase as a continuous ODE and by integrating it within the manifold of low-rank matrices. The low-rank subnetworks and their ranks are determined and adapted during the training phase, allowing the overall time and memory resources required by both training and inference phases to be reduced significantly. We will illustrate the efficiency of this approach on a variety of fully connected and convolutional networks.

The talk is based on:
S Schotthöfer, E Zangrando, J Kusch, G Ceruti, F Tudisco
Low-rank lottery tickets: finding efficient low-rank neural networks via matrix differential equations
(to appear on NeurIPS 2022)

Speakers: Andrey Kiselev, AASS, Örebro University and Stina Liivamäe, AI Sweden

Date: Friday, November 4, at 13.15, in T213

Title: AI++: Autonomous Systems in Industrial Applications

Abstract: The term AI can have many different meaning in different contexts and this presentation is going to look at the application of various AI methods in industrial applications. We will look into several examples of previous and current projects of various size and see how those projects are implemented from ideas to active collaboration and possibly to market-ready products. We will also discuss how various research activities are supported on the university level and through national wide networks to increase the impact.

Speaker: François Rousse, Örebro University

Date: Friday, October 14, at 13.15, in T213

Title: Phase-space representation for fermionic Quantum Dynamics

Abstract: Phase-space representations for bosonic Quantum Dynamics were introduced in Quantum Optics (i.e. for photons) in the 60’s. Some more practical computational maturity was achieved first in the 80-90’s through the development of the positive-P representation and the Truncated Wigner Approximation (TWA), and through new numerical implementation of the corresponding stochastic equations.  Then experimental progress in the field of Bose-Einstein condensates motivated applications of phase-space representations for (massive) bosonic particles, such as cold dilute atoms around 2000.

The more general understanding of matter also requires investigations of fermionic particles and of electronic structures, i.e. there is a need to develop simulation methods for fermions.  Phase-space methods maps the Hamiltonians to multidimensional Partial Differential Equations (PDE) for a quasi-probability density representing the Quantum Dynamics in an overcomplete basis. The PDEs are then mapped to stochastic differential equations (SDE) for stochastic sampling of physical observables. This last step is a standard procedure but allows for different types of so-called gauge degrees of freedoms, that can be explored to increase the practical numerical performance of the methods.  

We have explored a new phase-space representation for fermions, the fermionic Truncated Wigner Approximation (fTWA), first presented in 2017. fTWA is an approximative method which we have shown have major advantages over the standard mean-field method, in capturing higher order correlations of the quantum states. The first fTWA method have been explored on a large hexagonal 2D lattice, resembling the electronic structure of graphene. While the latter exact numerical-matrix-equation-based simulation have only been carried out on minimalistic 1D systems so far. 

In addition we have worked on extending the useful simulation time for another exact phase-space representation for fermions called the Gaussian phase-space representation (GPSR). We have done this by building in the mentioned gauge freedom into numerical matrices, obeying a constrained matrix equation. 

Speaker: Andrii Dmytryshyn, Örebro University

Date: Friday, September 16, at 13.15, Zoom meeting

Title: Versal deformations of matrices

Abstract: Jordan canonical form for matrices is well known and studied with various purposes but reduction to this form is an  unstable operation: both the corresponding canonical form and the reduction transformation depend discontinuously on the entries of an original matrix. This issue complicates the use of the canonical form for numerical purposes. Therefore V.I. Arnold introduced a normal form to which an arbitrary family of matrices A' close to a given matrix A can be reduced by similarity transformation smoothly depending on the entries of A’. He called such a normal form a versal deformation of A. 

In this presentation we will discuss versal deformations and their use in investigation of possible changes in canonical forms (eigenstructures), reduction of unstructured perturbations to structured perturbations, and codimension computations. 

Speaker: Fernando De Terán, Universidad Carlos III de Madrid

Date: Friday, May13 , at 13.15, Zoom meeting

Title: On the consistency of the matrix equation X^TAX=B when B is either symmetric or skew

Abstract: In this talk, we analyze the consistency of the matrix equation X^T AX=B, (1) where A in an mxm complex matrix, X in an mxn complex matrix (unknown), and B (complex nxn) is either symmetric or skew-symmetric (and (·)^T means the transpose). In particular, we first provide a necessary condition for (1) to have a solution X. Then, we will prove that this condition is also sufficient for most matrices A and an arbitrary symmetric (or skew) matrix B. We want to emphasize that the question on the consistency of (1), when B is symmetric (respectively, skew), is equivalent to the following problem: given a bilinear form over C^m (represented by the matrix A), find the maximum dimension of a subspace such that the restriction of the bilinear form to this subspace is a symmetric (resp., skew) non-degenerate bilinear form.

Speaker: Johan Hellsvik, KTH

Date: Friday, April 1 , at 13.15, Zoom meeting

Title: The Dardel HPE Cray EX supercomputer at PDC

Abstract: The CPU partition of the HPE Cray EX supercomputer Dardel has now entered regular operation. In this talk I will give an introduction to Dardel, the Cray programming environment, and the experiences learned so far on how to obtain good performance for scientific codes. An overview will be given on some of the scientific application programs that are now running on Dardel and how these how been tuned for AMD EPYC CPUs. In the spring the hardware for the GPU partition will be delivered to PDC. The GPU nodes with AMD Instinct MI250X GPUs will provide the main part of the total performance of Dardel. Examples will be given on some of the software porting activities that are currently ongoing to get codes ready for the GPU partition.

Speaker: Alan Edelman, Massachusetts Institute of Technology

Date: Friday, March 4, at 13.15, Zoom meeting

Title: Generalizing Orthogonal Matrices: On the Structure of the Solutions to the Matrix Equation G*JG=J

Abstract: We study the mathematical structure of the solution set to the matrix equation G*JG=J for a given square matrix J. In the language of pure mathematics, this is a Lie group which is the isometry group for a bilinear (or a sesquilinear) form.

We found that on its own, the related (tangent space) equation X*J+JX= 0 is hard to solve. By throwing into the mix the complementary linear equation X*J−JX= 0, we find that rather than increasing the complexity, we reduce the complexity.

We explicitly demonstrate computation of solutions, visualizations, and closure hierarchies that connect to previous work by Dmytryshyn, Futorny, Kågström, Klimenko, and Sergeichuk.

Joint work with Sung Woo Jeong.

Speaker: Jens Fjelstad, Örebro University

Date: Friday, February 4, at 13.15, Zoom meeting

Title: On quantum representations of mapping class groups from a finite group

Abstract: Topological quantum field theory (TQFT) is a gadget that produces two kinds of data, topological invariants and finite dimensional representations of mapping class groups of surfaces, so called quantum representations. These data are of interest both for Mathematics and Physics, one recent application being in quantum computing. One construction of TQFT is via a certain Hopf algebra constructed from a finite group, producing the quantum representations relevant for this talk. I will briefly introduce the ingredients such as mapping class groups of surfaces and the class of quantum representations associated to a finite group, and then discuss work (partly in collaboration with Jürgen Fuchs) to determine their structure in terms of finite group data. Recent results are focused on the restriction of a quantum representation to the terms in a filtration of a mapping class group known as the Johnson filtration.

Speaker: David Cohen, Chalmers University of Technology

Date: Friday, December 10, at 13.15, Zoom meeting

Title: Efficient discretisations of stochastic Hamiltonian and Poisson systems

Abstract: We start by recalling classical results on time discretisations of (deterministic) Hamiltonian and Poisson systems. We will then randomly perturbed such systems, present and analyse various time integrators for an efficient simulations of stochastic Hamiltonian and Poisson systems.

The presentation is based on joint works with C-E. Bréhier, C. Chen, R. D'Ambrosio, K. Debrabant, T. Jahnke, A. Lang, A. Rößler and G. Vilmart.

Speaker: Jens Fjelstad, Örebro University

Date: Friday, November 12, at 13.15 in T213

Title: On quantum representations of mapping class groups from a finite group

Abstract: Topological quantum field theory (TQFT) is a gadget that produces two kinds of data, topological invariants and finite dimensional representations of mapping class groups of surfaces, so called quantum representations. These data are of interest both for Mathematics and Physics, one recent application being in quantum computing. One construction of TQFT is via a certain Hopf algebra constructed from a finite group, producing the quantum representations relevant for this talk. I will briefly introduce the ingredients such as mapping class groups of surfaces and the class of quantum representations associated to a finite group, and then discuss work (partly in collaboration with Jürgen Fuchs) to determine their structure in terms of finite group data. Recent results are focused on the restriction of a quantum representation to the terms in a filtration of a mapping class group known as the Johnson filtration.

Speaker: Massimiliano Fasi, Durham University, UK

Date: Friday, October 15, at 13.15, Zoom meeting

Title: CPFloat: A C Library for Emulating Low-Precision Arithmetic

Abstract: Low-precision floating-point arithmetic can be simulated via software by executing each arithmetic operation in hardware and rounding the result to the desired number of significant bits. For IEEE-compliant formats, rounding requires only standard mathematical library functions, but handling subnormals, underflow, and overflow demands special attention, and numerical errors can cause mathematically correct formulae to behave incorrectly in finite arithmetic. Moreover, the ensuing algorithms are not necessarily efficient, as the library functions these techniques build upon are typically designed to handle a broad range of cases and may not be optimized for the specific needs of floating-point rounding algorithms. CPFloat is a C library that offers efficient routines for rounding arrays of binary32 and binary64 numbers to lower precision. The software exploits the bit level representation of the underlying formats and performs only low-level bit manipulation and integer arithmetic, without relying on costly library calls. In numerical experiments the new techniques bring a considerable speedup (typically one order of magnitude or more) over existing alternatives in C, C++, and MATLAB. To the best of our knowledge, CPFloat is currently the most efficient and complete library for experimenting with custom low-precision floating-point arithmetic available in any language.

Speaker: Hugo U.R. Strand, Örebro University

Date: Friday, September 17, at 13.15, Zoom meeting

Title: Hands-on high-order orthogonal-polynomial methods for an integrodifferential equation

Abstract: High-order orthogonal-polynomial approximation methods and their application is still an evolving field in numerical mathematics [1]. Recent advances has enabled extreme high-order approximations [2] and new algorithmic developments has opened the door for applications on integral equations [3].

In this talk we will present an application of these methods to the integrodifferential "Dyson" equation, that appears in the context of perturbation theory in many-body quantum physics [4].

[1] S. Olver, R.M. Slevisksy, A. Townsend, Acta Numerica pp. 573-699 (2020)
[2] I. Boagert, SIAM J. Sci. Comput., v36 no. 3 pp. A1008-A1026 (2014)
[3] N. Hale, A. Townsend, SIAM J. Sci. Comput., v36 no. 3 pp. A1207-A1220 (2014)
[4] X. Dong, D. Zgid, E. Gull, H.U.R. Strand, J. Chem. Phys. 152, 134107 (2020)

Speaker: Attila Szilva, Uppsala University

Date: Friday, May 7, at 13.15, Zoom meeting

Title: From elections to science of cities: an introduction to complexity

Abstract: Animals from rats to the blues whales are built up from cells arranged in networks. The topology of the underlying networks explains the so-called Kleiber’ scaling law, which states that an animal's metabolic rate scales to the ¾ power of the animal's mass (a cat having a mass 100 times that of a mouse will consume only about 32 times the energy the mouse uses). This scaling is sublinear because the power is less than 1. In the presentation, it will be shown that the infrastructure of cities (the length of electric cables or the number of gas stations) also follows universal sublinear scaling law while in socio-economic dimensions (GDP per capita, innovation, crime) cities are superlinear. They are as a result of the individual interactions proven by a large set of mobile phone data. The concept of scaling and universality is originated in statistical physics where a large complex system emerges from simple interactions, and its behavior is almost totally independent of its microscopic structure. Another examples are the democratic elections for which physics inspired models will be also discussed in the presentation.

The talk is based mostly on the books of Geoffrey West: Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life in Organisms, Cities, Economies, and Companies and Stefan Thurner, Rudolf Hanel and Peter Klimek: Introduction to the theory of complex systems and the paper Social influence with recurrent mobility and multiple options by Jerome Michaud and Attila Szilva published in Physical Review E 97 (6), 062313.

Speaker: Jakob Palmkvist, Örebro University

Date: Friday, April 9, at 13.15, Zoom meeting

Title: Infinite-dimensional Lie superalgebras

Abstract: I will review how the classification of simple finite-dimensional Lie algebras leads to the construction of contragredient Lie algebras, which in general are infinite-dimensional, and how these can be further generalised to contragredient Lie superalgebras. I will then explain how a modification of the construction gives rise to a new class of non-contragredient Lie superalgebras, called tensor hierarchy algebras, whose finite-dimensional members are the simple Lie superalgebras of Cartan type. Tensor hierarchy algebras have proven useful in describing gauge structures in physical models related to string theory. I will describe some of the remarkable features they exhibit.

Speaker: Maria ElGhaoui, Sorbonne University, Paris (France) and Saint Joseph University, Beirut (Lebanon)

Date: Friday, March 5, at 13.15, Zoom meeting

Title: A Trefftz Method With Reconstruction of The Normal Derivative Applied to Elliptic Equations

Abstract: There are many classical numerical methods for solving boundary value problems on general domains. The Trefftz method is an approximation method for solving linear boundary value problems arising in applied mathematics and engineering sciences. This method consists to approximate the exact solution through a linear combination of trial functions satisfying exactly the governing differential equation.  One of the advantages of this method is that the number of trial functions per cell is O(m), asymptotically much less than the quadratic estimate O(m^2) for finite element and discontinuous Galerkin approximations.  

For a Laplace model equation, we present a high order Trefftz method with quadrature formula for calculation of normal derivative at interfaces.   We introduce a discrete variational formulation and study the existence and uniqueness of the discrete solution.   A priori error estimate is then established and finally, several numerical experiments are shown

Speaker: Mårten Gulliksson, Örebro University

Date: Friday, February 5, at 13.15, Zoom meeting

Title: My Least Squares Problems

Abstract: During a period of more than 30 years I have now and then worked with different kinds of least squares problems. These include simple linear least squares in different settings but also much more abstract ill-posed nonlinear problems in Hilbert spaces. Some of the problems were applied such as, e.g., surface fitting, neural networks, and estimating elastic properties in paper. Others were more theoretical in nature involving convergence aspects of methods, perturbation theory, and sparsity. The talk will end with some open least squares problems.

Speaker: Andrii Dmytryshyn, Örebro University

Date: Friday, November 11, at 13.15 in T211

Title: Recovering a perturbation of a matrix polynomial from a perturbation of its linearization

Abstract: A number of theoretical and computational problems for matrix polynomials are solved by passing to linearizations. Therefore, a perturbation theory results for the linearizations need to be related back to matrix polynomials. We present an algorithm that finds which perturbation of matrix coefficients of a matrix polynomial corresponds to a given perturbation of the entire linearization pencil. We consider general matrix polynomials as well as polynomials of odd grade whose coefficients are (skew) symmetric. Moreover, we find transformation matrices that, via strict equivalence or, respectively, congruence, transform a perturbation of the linearization to the linearization of a perturbed polynomial.

Speaker: Massimiliano Fasi, Örebro University

Date: Friday, October 9, at 13.15 in T207

Title: Generating extreme-scale matrices with specified singular values or condition number

Abstract: The randsvd matrix is a widely used test matrix constructed as the product A = USV, where U and V are random orthogonal or unitary matrices from the Haar distribution and S is a diagonal matrix of singular values. Such matrices are random but have a specified singular value distribution. Forming an m-by-n randsvd matrix requires a number of floating-point operations cubic in both m and n, which is prohibitively expensive at extreme scale. Moreover the randsvd construction requires a significant amount of communication, making it unsuitable for distributed memory environments. By dropping the requirement that U and V be Haar distributed and that both be random, we derive new algorithms for forming A that have cost linear in the number of matrix elements and require a low amount of communication and synchronisation. We specialise these algorithms to generating matrices with specified 2-norm condition number. Numerical experiments show that the algorithms have excellent efficiency and scalability.

Speaker: Malte Litsgård, Uppsala University

Date: Friday, September 11, at 13.15 in T207

Title: Degenerate Kolmogorov-type equations with rough coefficients - Potential theory and boundary regularity

Abstract: Today Kolmogorov-type operators with low-regularity coefficients have applications in many different areas: analysis, physics, and finance, to name a few. In particular the study of local regularity of weak solutions to PDEs associated to such operators is relevant in for example kinetic theory, where these equations appear in relation to the study of conditional regularity of the Boltzmann and Landau equations. In this talk I will give some background for these operators and present some results concerning potential theory and boundary regularity in Lipschitz-type domains.

Speaker: Magnus Ögren, Örebro universitet

Date: Friday, March 6, at 13.15 in T1210

Title: A numerical damped oscillator approach to constrained Schrödinger equations

Abstract: This article explains and illustrates the use of a set of coupled dynamical equations, second order in a fictitious time, which converges to solutions of stationary Schr\"{o}dinger equations with additional constraints.

We include three qualitative different numerical examples: the radial Schr\"{o}dinger equation for the hydrogen atom; the two-dimensional harmonic oscillator with degenerate excited states; and finally a non-linear Schr\"{o}dinger equation for rotating states.

The presented method is intuitive, with analogies in classical mechanics for damped oscillators, and easy to implement, either in own coding, or with software for dynamical systems.

Hence, we find it suitable to introduce it in a continuation course in quantum mechanics or generally in applied mathematics courses which contain computational parts.

Speaker: Mikael Hansson, Örebro universitet

Date: Friday, February 7, at 13.15 in T1210

Title: An introduction to Coxeter groups, the word problem, and twisted involutions

Abstract: Given a group generated by an alphabet S, the word problem is the problem of deciding whether two words with letters from S represent the same group element. For the important class of Coxeter groups, there is a solution based on the so-called word property, which will be described. I shall motivate and define Coxeter groups, so no knowledge of these groups is assumed. Then, twisted involutions will be introduced. We shall see that in this case, too, there is a word property which solves the word problem.

This talk is based on joint work with Axel Hultman, Linköping University.

Speaker: Per Enflo

Date: Friday, January 10, at 13.15 in T1210

Title: Några aktuella problemområden inom funktionalanalysen och dess tillämpningar

Abstract: 1/ För 200 år sedan visade Fourier att en godtycklig periodisk funktion kan delas upp och skrivas som en summa (linjär kombination) av sinus- och cosinusfunktioner, som alltså utgör ett system av genererande funktioner.  Men sinus- och cosinusfunktioner kan inte alltid användas och idag bedrivs mycket forskning  att hitta lämpliga  system av genererande funktioner (baser) i många olika sammanhang, såväl inom ”ren” matematik som inom dess tillämpningar.

2/ Vid studiet av nxn-matriser är det ofta viktigt att finna matrisernas egenvektorer och deras egenvärden. Men för generaliseringen av matriser till en oändligt-dimensionell situation (de kallas då linjära operatorer) så finns inte alltid egenvektorer.  Men kanske finns det ”invarianta underrum”
(en svagare men ändå viktig egenskap). Det nu drygt 80 år gamla problemet, huruvida operatorer på Hilbertrummet alltid har invarianta underrum, är nog det mest berömda olösta problemet inom funktionalanalysen.

Speaker: Danny Thonig, Örebro University.

Date: Wednesday, December 4, at 15.00 in T217

Title: An introduction to ab-initio magnetization dynamics

Abstract: The time-integrated amount of stored information is doubled roughly every eighteen months, and since the majority of the world’s information is stored in magnetic media, the possibility to write and retrieve information in a magnetic material at ever greater speed, bigger data transmission rates and with lower energy consumption, has obvious benefits for our society. Hence the seemingly simple switching of a magnetic unit, a magnetic bit, is a crucial process which defines how efficiently information can be stored and retrieved from a magnetic memory. From an application point of view, it is apparent that it is advantageous to be able to switch the magnetization of a bit as fast as possible while minimizing energy losses. It implies also to understand the microscopic origin of the magnetization dynamics in magnetic materials.

Within my talk I will give a general introduction about what magnetism is, from where it originates, and how it can be treated in modeling. Why magnetism is materials specific and - although known as rather static - magnetism is dynamic will be motived after the introduction part. Here, the equation of motion will be discussed more in detail and applied to certain questions that occurred from experimental observations. All this is implemented in the software package "Cahmd" (, which was developed by the PI. I will finish my presentation with an outlook on future challenges.

Speaker: Zhaojun Bai, University of California, Davis.

Date: Tuesday, November 26, at 15.00 in T215

Title: Rayleigh quotient optimizations and eigenvalue problems

Abstract: Many computational science and data analysis techniques lead to optimizing Rayleigh quotient (RQ) and RQ type objective functions, such as computing excitation states (energies) of electronic structures, robust classification to handle uncertainty and constrained data clustering to incorporate a prior information. We will discuss origins of recently emerging RQ optimization problem, variational principles, and reformulations to algebraic linear and nonlinear eigenvalue problems. We will show how to exploit underlying properties of eigenvalue problems for designing eigensolvers, and illustrate the efficacy of these solvers in applications. 

Speaker: Per Bäck, Mälardalen University

Date: Friday, November 8, at 13.15 in T1210

Title: Theory of non-commutative and non-associative polynomial rings

Abstract: In this talk, I will give an introduction to the theory of non-commutative polynomial rings known as Ore extensions, and then show how this can be generalized to the non-associative setting. In particular, I will discuss hom-associative Ore extensions and how they provide a framework for deforming otherwise rigid algebras appearing in e.g. quantum physics, such as the Weyl algebras. I will also show how these deformations induce deformations of the corresponding commutator Lie algebras into so-called hom-Lie algebras.

In the end, we will see how the classical Hilbert’s basis theorem for commutative polynomial rings can be extended to a non-commutative and non-associative version, and applications thereof including quaternionic and octonionic rings.

Speaker: Anders Tengstrand

Date: Wednesday, October 30, at 13.15 in T217

Title: Det oändligt stora och det oändligt lilla

Abstract: Det oändligt stora och det oändligt lilla har i alla tider utmanat matematiker.  Med hjälp av dessa begrepp har man skapat effektiva verktyg för att lösa problem inom matematiken och dess tillämpningar. Samtidigt har matematiker som använt sig av oändligt små och oändligt stora tal utsatts för kritik för otydlighet och brist på konkretion. Jag ska i min föreläsning belysa denna problematik med exempel från antiken fram till 1900-talet.

Speaker: Johan Andersson (Örebro Universitet)

Date: Friday, October 11, at 13.15-14.30 in T2102

Title: Universality of zeta and L-functions

Abstract: Voronin proved in the seventies that any zero-free analytic function f(s) on a disc |s-3/4|<r<1/4 which is continuous up to its boundary may be uniformly approximated as closely as desired by imaginary shifts of the Riemann zeta-function. We say that the Riemann zeta-function is universal. In the last 40 years this property has been proved for a wide range of zeta and L-functions, such as automorphic L-functions and the Selberg zeta-function. I will talk about some of my recent work in the field. In particular I have recently proved (arXiv:1809.03444) that the Euler-Zagier multiple zeta-function is universal in several complex variables, thus giving the first example of a Dirichlet series that is universal in more than one complex variable. I will also talk about recent work in progress where I give the first universality theorem for the Hurwitz zeta-function with an algebraic irrational parameter.

Speaker: Mac Panahbehagh (Örebro Universitet)

Date: Friday, September 13, at 13.15-14.30 in T1210

Title: Simulations of a porous particle settling in density stratified ambient 

Abstract: We study numerically the settling of a porous sphere in a density-stratified ambient fluid. Simulations are validated against prior laboratory experiments and compared to two mathematical models. Two main effects cause the particle to slow down as it enters a density gradient: lighter fluid within the particle and entrainment of the density-stratified ambient fluid. The numerical simulations accurately capture the particle retention time. We quantify the delay in settling due to ambient fluid entrainment and lighter internal fluid becoming denser through diffusion as a function of different parameters. A simple fitting formula is presented to describe the settling time delay as a function of each of those non-dimensional parameters.

Speaker: Simon Streib (Uppsala Universitet)

Date: Wednesday, August 21, at 13.15-14.30 in T1210

Title: The Barnett/Einstein-de Haas effects and magnetoelastic coupling in magnetic insulators

Abstract: The Barnett effect is the magnetization induced by mechanical rotation of an uncharged body while the Einstein-de Haas effect is its reciprocal: rotation induced by magnetization. I give an historical overview of these effects, including a discussion of the recent observation of the nuclear Barnett effect [1], and present a simple derivation of the Barnett effect based on the conservation of energy and angular momentum. Finally, I discuss the physical origin of the magnetoelastic coupling in magnetic insulators and its effect on the lifetime of magnetic excitations [2].

[1] M. Arabgol and T. Sleator, Phys. Rev. Lett. 122, 177202 (2019).

[2] S. Streib, N. Vidal-Silva, K. Shen, and G. E. W. Bauer, Phys. Rev. B 99, 184442 (2019).

Speaker: Froilán M. Dopico (Universidad Carlos III de Madrid)

Date: Wednesday, May 29, at 15:15 in T1210

Title: Local linearizations of rational matrices with application to nonlinear eigenvalue problems

Abstract: The numerical solution of nonlinear eigenvalue problems (NLEP) has attracted considerable attention since 2004, mainly as a consequence of the influential reference "Nonlinear eigenvalue problems: a challenge for modern eigenvalue methods" (GAMM Mitt. Ges. Angew. Math. Mech., 2004) by V. Mehrmann and H. Voss. A variety of methods have been developed for these problems and the most successful ones can be found in the recent survey "The nonlinear eigenvalue problem" (Acta Numer., 2017) by S. G\"{u}ttel and F. Tissuer. Both for dense-medium-size and for large-scale problems the preferred methods consist of three steps: 

(1) to approximate the NLEP by a rational eigenvalue problem (REP) in a certain region; 

(2) to construct a linear eigenvalue problem (LEP) that has the same eigenvalues of the REP in the region of interest; 

(3) to compute via the QZ method for dense problems or via structured rational Krylov methods for large-scale problems the eigenvalues of the LEP. 

The purpose of this talk is to develop a mathematical local theory of linearizations of REPs that allows us, among other things, to establish rigorously the properties of the LEPs that have been used for solving NLEPs in a number of recent references.

Speaker: Massimiliano Fasi (University of Manchester)

Date: Tuesday, May 21, at 15:15 in T1210

Title: Substitution algorithms for rational matrix equations

Abstract: Functions of matrices defined as solutions to matrix equations play an important role in many applications. As rational approximation is a customary tool in algorithms for evaluating this class of functions, one may wonder whether it be possible to solve numerically equations of the form $r(X)=A$, where $A$ and $X$ are square matrices and $r$ is a rational function.

As it turns out, accurate and efficient techniques for this problem can be derived by inverting, through a substitution strategy, computational schemes for evaluating rational matrix functions. The resulting methods all exploit the Schur decomposition of the input matrix to reduce the problem to upper triangular form, and for triangular matrices they yield the same computational cost as the evaluation schemes from which they are obtained. This suggests that solving rational matrix equations is not more difficult than evaluating rational functions at a matrix argument.

These methods can be used in a natural way as building blocks in algorithms for computing functions of matrices defined via matrix equation of the type $f(X) = A$, where $f$ is a primary matrix function.