Department of Mathematics and Systems Analysis


Lectures, seminars and dissertations

* Dates within the next 7 days are marked by a star.

Atte Mäkelä
(Master thesis talk) Deep reinforcement learning as a tool for search engine campaign budget optimization
* Wednesday 28 July 2021,   10:00,   Zoom

Past events

Miikka Runolinna
Confidence interval estimation of extreme quantiles in semiconductor processes (Master's thesis presentation)
Thursday 15 July 2021,   14:15,   Zoom
Further information

Anton Vavilov
Planar Qasiconformal Mappings: Fundamental Properties and Characterizations
Friday 09 July 2021,   12:00,   Zoom
Further information
Master's thesis presentation

Axel Päivänsalo
Robust PCA using the L1-norm (Bachelor thesis talk)
Wednesday 07 July 2021,   11:00,   Zoom

Henry Pietilä
Height Estimation Using Sensor Fusion (Bachelor thesis talk)
Thursday 01 July 2021,   13:00,   Zoom

Dr. Negin Karimi
Machine learning talk series: Distributed machine learning
Wednesday 23 June 2021,   15:15,   Zoom
Further information
Nowadays, with the growing technological expansion of the world, there is a huge amount of data that is created every day. Currently, one of the best approaches for storing data is distributed systems. The earliest example of distributed storage systems was invented in 1970 when for the first time computers would be able to send messages to other systems with a local IP address. Telephone and cellular networks are other examples of distributed systems. These systems perform machine learning tasks in a distributed environment. In this talk, I'll present an introduction to distributed machine learning, which refers to multi-node machine learning algorithms and systems that are designed to improve performance, increase accuracy, and scale to larger input data sizes.
ANTA Seminar

Iiro Tynnilä
Kvasi- Monte Carlo -menetelmät
Wednesday 23 June 2021,   10:45,   Zoom
Further information

Paul Närhi
Lämpöyhtälön vahvan maksimiperiaatteen todistus
Wednesday 23 June 2021,   10:00,   Zoom
Further information

Dr. Taoufiq Damir
Lattice-based cryptography, part 2 (talk series for PhD students in number theory)
Thursday 17 June 2021,   15:00,   Zoom
Further information
ANTA Seminar

Dr. Taoufiq Damir
Lattice-based cryptography, part 1 (talk series for PhD students in number theory)
Thursday 10 June 2021,   15:00,   Zoom
Further information
ANTA Seminar

Prof. Guillermo Mantilla-Soler
New perspectives on Arithmetic equivalence
Wednesday 09 June 2021,   15:15,   Zoom
Further information
Two number fields are called arithmetically equivalent if their Dedekind zeta functions coincide. It is known that two arithmetically equivalent number fields share degree, discriminant, Galois closure and several other invariants. In this talk I'll recall briefly all these classical results and show how the a perspective of Galois representations and one from arithmetic geometry lead us to prove new results, and put all previous in a clear context, about A.E.
ANTA Seminar

Dr. Taoufiq Damir
On lattice constructions (talk series for PhD students in number theory)
Thursday 27 May 2021,   15:00,   Zoom
Further information
ANTA Seminar

Max Cedercreutz
Wind comfort analysis in forested areas (Master's thesis presentation)
Friday 21 May 2021,   11:00,   Zoom
Further information

Pekka Huhtala
Loop O(n) models: a numerical transfer matrix study of long range order and s-holomorphicity
Thursday 20 May 2021,   15:15,   Zoom
Further information
Master's thesis presentation

Dr. Taoufiq Damir
Lattices and wiretap channels (talk series for PhD students in number theory)
Thursday 20 May 2021,   15:00,   Zoom
Further information
ANTA Seminar

Jere Karttunen
The effect of transducer curvature on the required energy in multi-locus transcranial magnetic stimulation (Bachelor thesis talk)
Wednesday 19 May 2021,   13:00,   Zoom

Dr. Razane Tajeddine (HIIT)
Machine learning talk series: Privacy preserving data sharing on vertically partitioned data
Wednesday 12 May 2021,   15:15,   Zoom
Further information
Predictive machine learning uses big amounts of data to predict the likelihood of future outcomes. To build predictive models, confidential personal data needs to be collected and shared. Therefore, user data privacy is a major concern when training such models. In many cases, the data is partitioned between multiple servers or parties, where each party holds parts of the data which can not be revealed to the other parties due to privacy concerns. Combining data from different parties gives additional information, and thus the quality of the model is significantly improved. Applying machine learning algorithms to a partitioned dataset without violating the privacy of the users is a challenging task. This talk will be concerned with methods for building predictive models on partitioned data while preserving user privacy.
ANTA Seminar

Prof. Henrik Garde (Aarhus University)
Series reversion in Calderón's problem
Wednesday 12 May 2021,   13:15,   Zoom
Further information
I will talk about reconstruction in the inverse problem of electrical impedance tomography (Calderón's problem), on finding a conductivity distribution from boundary electrical measurements. Here I introduce a series reversion of the forward map, that leads to higher order numerical methods for reconstructing a perturbation to a conductivity coefficient, without increasing the computational complexity compared to solving a linearised problem. The numerical methods can be derived for both a continuum model and for a (smoothened) complete electrode model. Joint work with Nuutti Hyvönen

Dr. Taoufiq Damir
Introduction to lattices (talk series for PhD students in number theory)
Thursday 06 May 2021,   15:00,   Zoom
Further information
ANTA Seminar

Dr. Laura Jakobsson
Machine learning talk series: Machine learning meets commutative algebra: Table ideal identification
Wednesday 05 May 2021,   15:15,   Zoom
Further information
Table ideals are monomial ideals that come from tables, i.e. arrangements of integers that satisfy some conditions on the entries, and these ideals have many nice properties. Given a table, the ideal coming from it in a non-minimal form is easy to compute. However, the other direction is not as easy, that is, given a monomial ideal how do we know if it is a table ideal? In this talk, I will introduce the machine learning problem of distinguishing table ideals from non-table ideals, and the results we have obtained from using machine learning to this maths question.
ANTA Seminar

Camilla Hollanti
Wireless communications and number field lattice codes (talk series for PhD students in number theory)
Thursday 29 April 2021,   15:00,   Zoom
Further information
ANTA Seminar

Dr. Mika Malinen ( Aalto University)
A local reparametrization by approximate lines of curvature coordinates
Thursday 29 April 2021,   13:15,   Zoom
Further information
A surface parametrization by orthogonal curvilinear coordinates offers many benefits from the viewpoint of both analysis and practical implementation. However, ways to construct such representations are not immediate. The utility of lines of curvature as a strategy for obtaining such coordinates globally in a smooth surface is severely limited by the presence of umbilical points. Nevertheless, a global parametrization is not necessarily needed in the numerical solution of PDE problems on surfaces, which gives us a motivation for developing methods to obtain a local reparametrization by consistently accurate lines of curvature coordinates. A way to avoid the possible instabilities of reparametrization in a vicinity of an isolated umbilical point is considered via a sensitivity analysis. Some illustrative examples are also shown.

Dr. Laia Amoros
Machine learning talk series: Practical introduction to machine learning for mathematicians
Wednesday 28 April 2021,   15:15,   Zoom
Further information
This talk is aimed at mathematicians that have heard about machine learning and are interested in getting more familiar with the field, but don’t know where to start. Machine learning (ML) covers a huge amount of algorithms that aim at learning useful information from data. Whether the best learning option is supervised, unsupervised or any other kind of learning depends on the type of our data. It turns out that handling the data takes most of the effort, and defining a model to make predictions out of the data can be done with a few simple lines of Python. We will see some easy examples in a Python notebook environment, so the interested mathematician can reproduce the experiments with their own data. We will show how ML can in particular be used to learn information about mathematical objects, which in turn can be useful to better know your data, be it monomial ideals, number fields or Calabi-Yau hypersurfaces. For the slides and related material, click "further info".
ANTA Seminar

Samuel Nurmi
(Bachelor Thesis Talk) Controlling coil overlap in multi-locus transcranial magnetic stimulation transducers
Thursday 22 April 2021,   13:00,   Zoom
Further information
Bachelor thesis talk

Stavros Evdoridis
On geometric properties of harmonic and polyharmonic mappings
Wednesday 14 April 2021,   12:00,   Zoom
Link to the event

Tuomas Leidén
Predictive Multi-target Learning for Industrial Plant Engineering (Master's thesis presentation)
Thursday 08 April 2021,   13:00,   Teams
Further information

Prof. Paul Van Dooren, Université catholique de Louvain
Detecting roles in large graphs
Thursday 01 April 2021,   13:15,
Computing meaningful clusters of nodes is crucial to analyze very large networks, but since this is a combinatorial optimization problem, its complexity can be prohibitive. In this talk, we present a pairwise node similarity measure that allows to extract roles, i.e. group of nodes sharing similar flow patterns within a network. We then propose a low rank iterative scheme to approximate the similarity measure for very large networks and we show that our low rank similarity score successfully extracts the different roles in random graphs of Erdos-Renyi type. We then argue that when projecting our similarity measure onto a low rank manifold (say, of rank r), the role models can be recovered using classical clustering techniques in a r-dimensional space, which reduces the complexity significantly. We illustrate these ideas with a number of real-world examples.

Perttu Yli-Opas
Detumble algorithm for the AuroraSat-1 propulsive attitude control system
Thursday 25 March 2021,   14:15,
Bachelor's thesis talk

Luca Ferranti (University of Vaasa)
Optimised polynomial solvers for computer vision and positioning
Monday 22 March 2021,   10:15,
Systems of polynomial equations naturally arise in several engineering applications. General polynomial solving techniques may not be enough to solve these problems, especially at higher robustness and efficiency requirements. For this reason, specialised polynomial solvers, exploiting the specific structure of the problem, are desired. In the talk, I will describe how optimised polynomial solvers can be automatically generated using Gröbner bases. I will also give an overview of several success stories of these techniques in computer vision and positioning applications, where these automatically generated solvers could solve challenging problems in microseconds. Finally, I will briefly discuss what problems are still open and what are current research trends in this area.
Algebra and Discrete Mathematics Seminar

Lauri Nyman
Solving Systems of 3 Equations in 3 Variables via Resultants
Thursday 18 March 2021,   13:15,
This talk is a presentation of my master's thesis. I will discuss a root-finding algorithm that computes numerically the real solutions of a system of three equations in three variables by using the Cayley resultant method. The Cayley resultant method works by first approximating the system of functions by polynomials, and then transforming the resulting multidimensional polynomial root-finding problem into a polynomial eigenvalue problem. The resulting polynomial eigenvalue problem is solved through a linearization given by the colleague pencil. I will present the details of the implementation as well as numerical results.

Olli Winberg
Compressed Sensing of Fourier sparse quasiparticle-interference map
Wednesday 10 March 2021,   11:00,   Zoom:
Bachelor's thesis talk

Emilia Blåsten (Aalto University)
Inverse spectral problem on discrete graphs
Thursday 04 March 2021,   13:15,
I will introduce the results of my latest publication, arXiv:2101.10026, and its background. The inverse spectral problem in partial differential equation asks for determining unknown coefficients of given partial differential equation, or the underlying manifold structure, from boundary spectral measurements. These are the pairs of all eigenvalues with the boundary normal derivatives of their corresponding Dirichlet eigenfunctions to a fixed partial differential operator. The eigenvalues themselves are not enough to determine the underlying geometry: one cannot hear the shape of a drum. The question has been studied extensively for quantum graphs, but with limited success except for trees. My latest publication takes this topic into the context of discrete graphs. Given a subset of vertices (call it the boundary), and boundary values of Neumann eigenvectors to a graph Laplacian, we show that the rest of the graph structure can be determined assuming that there is "enough" boundary vertices.

Jaakko Pere
On estimation of extreme probability contours under multivariate elliptical Fréchet distributions (Master's thesis presentation)
Wednesday 03 March 2021,   09:00,

Maija Vahteristo
On the effects of cervical cancer screening (Master's thesis presentation)
Friday 19 February 2021,   10:00,

María Quintana Ponce, Aalto University
Linearizations of Rational Matrices
Thursday 18 February 2021,   13:15,
Zoom-link : The numerical solution of rational eigenvalue problems (REPs) is getting a lot of attention. REPs appear directly from applications or they can be used for solving arbitrary nonlinear eigenvalue problems (NLEPs) by approximating the corresponding NLEP by a REP and, then, solving the REP. They also play a fundamental role in system and control theory. Nowadays, one of the most competitive methods for solving REPs is based on linearizations of the corresponding rational matrix. In this talk, we introduce a new notion of linearization for rational matrices, and we provide very simple criteria to determine when a linear polynomial matrix is one of these linearizations. As an application, we study the structure of linearizations constructed in the literature for rational approximations of NLEPs.

Leevi Lindgren
On Moment-Based Estimation of Extreme Probability Contours Under Ellipticity (Master's thesis presentation)
Wednesday 10 February 2021,   10:00,   in Teams

Tuomas Tuukkanen
Haar measures on locally compact groups (harmonic analysis workshop)
Friday 05 February 2021,   10:15,
Seminar on analysis and geometry

Timo Takala
The duality of Hardy functions and BMO functions (harmonic analysis workshop)
Friday 29 January 2021,   10:15,
Seminar on analysis and geometry

Laura Jakobsson
A survey of the Miyata-Proudfoot-Ramos category of graphs, 2
Friday 15 January 2021,   11:00,
Alexander Engström

Christoffer Lundström
Monday 21 December 2020,   14:15,
BSc thesis presentation. The presentation will be in Swedish.

Laura Jakobsson
A survey of the Miyata-Proudfoot-Ramos category of graphs, 1
Friday 18 December 2020,   11:00,

Kim Myyryläinen
Rubio de Francia extrapolation theorem (harmonic analysis workshop)
Wednesday 16 December 2020,   10:15,
Seminar on analysis and geometry

Napoleon Freitas Paajanen
Connection between Littlewood-Paley operators and Calderón-Zygmund integral kernels (harmonic analysis workshop)
Tuesday 15 December 2020,   14:15,
Seminar on analysis and geometry

Niklas Miller
Algebraic number theory: an analysis of well-rounded lattices (MSc thesis presentation)
Tuesday 15 December 2020,   14:00,

Lari-Matti Tuomala
Secure coded multi-party matrix multiplication (MSc thesis presentation)
Tuesday 15 December 2020,   13:00,

Lauri Särkiö
Boundedness of convolution type Calderón-Zygmund operators (harmonic analysis workshop)
Tuesday 15 December 2020,   10:15,
Seminar on analysis and geometry

Kristian Moring
Supercaloric functions to the singular parabolic p-Laplace equation in the supercritical case
Monday 14 December 2020,   14:15,
Seminar on analysis and geometry

Napoleon Freitas Paajanen
Sobolev characterization based on Marcinkiewicz integral transform (diploma thesis presentation)
Monday 14 December 2020,   10:15,
Seminar on analysis and geometry

Pasi Hakulinen
On the Backward Monte-Carlo Method with Application to Fast Ion Simulation in Magnetized Plasma
Friday 11 December 2020,   14:15,

Henri Södergård (Aalto University)
Image deblurring by numerical linear algebra (BSc thesis)
Wednesday 09 December 2020,   14:15,
When taking a photograph, we usually want the resulting image to be a faithful reconstruction of the reality we see with our own eyes. However, almost all images will inevitably suffer from some form of imperfections such as blurriness. To fix this, we can deblur the image employing methods and techniques from numerical linear algebra to mathematically. This way, we can restore the maximum amount of information about the true image from a blurry image. In this talk, I plan to introduce some basic concepts in image deblurring, including the naive solution, the point spread function, spectral filtering (such as Truncated Singular Value Decomposition or the Tikhonov method) and regularisation methods (such as Generalised Cross Validation, the Discrepancy Principle and the L-Curve Criterion). I will also present the results of some experiments that I have run to compare some of these methods in practice, by testing them on some blurred images.
BSc Seminar

Lauri Hitruhin
Boundary behaviour of conformal and quasiconformal mappings
Monday 07 December 2020,   14:15,
Seminar on analysis and geometry

Cintia Pacchiano
Regularity properties for quasi-minimizers of a (p,q)-Dirichlet integral
Monday 07 December 2020,   10:15,
Seminar on analysis and geometry

Dr Jeta Molla (Aalto University)
Numerical methods for the stochastic wave equation
Tuesday 24 November 2020,   16:00,
Further information
The objective of this talk is to propose a full discretization for the stochastic wave equation. More specifically, the discontinuous Galerkin finite element method is used in space and analyzed in a semigroup framework, and an explicit stochastic position Verlet scheme is used for the temporal approximation. Numerical experiments illustrate our theoretical results on strong convergence rates. Further, we analyze and bound the expected energy and numerically show excellent agreement with the energy of the exact solution. Join Zoom Meeting Meeting ID: 689 1125 9210
Aalto Stochastics & Statistics Seminar / Leskelä

Emma-Karoliina Kurki
Integral estimates on Whitney chains
Monday 09 November 2020,   14:15,   M237
Seminar on analysis and geometry

Kim Myyryläinen
Median-type John–Nirenberg spaces
Monday 09 November 2020,   10:15,   M237
Seminar on analysis and geometry

Mohamed Taoufiq Damir
Well-rounded lattices and applications to physical layer security (Doctoral defence)
Friday 06 November 2020,   15:00,   Zoom
Further information

Julian Weigt
The dyadic maximal function has bounded variation
Monday 02 November 2020,   14:15,   M237
Seminar on analysis and geometry

Carlos Mudarra
Extensions of jets with smooth convex functions and applications
Monday 02 November 2020,   10:15,   M237
Seminar on analysis and geometry

Prashanta Garain
Nonexistence of variational minimizers related to a quasilinear singular problem in metric measure spaces
Monday 05 October 2020,   14:15,   M237
Seminar on analysis and geometry

Sauli Lindberg
Convex integration in magnetohydrodynamics
Monday 05 October 2020,   10:15,   M237
Seminar on analysis and geometry

Afrin Hossain
Fatou and Julia sets in Rational Iteration (kandiesitelmä)
Monday 21 September 2020,   13:15,   M237

Oona Oinonen
Dihedraalisen piilotetun aliryhmän ongelma (kandiesitelmä)
Monday 21 September 2020,   11:00,   Zoom

Antti Immonen
Secure Distributed Matrix Multiplication (BSc presentation)
Monday 21 September 2020,   10:15,   Zoom

Kalle Kytölä
Euler integrals and a quantum group
Monday 31 August 2020,   15:45,
mathematical physics end-of-summer seminar day (Kytölä)

Shinji Koshida
Point processes and fermionic algebras
Monday 31 August 2020,   15:00,
mathematical physics end-of-summer seminar day (Kytölä)

Eveliina Peltola
On large deviations of multiple SLEs
Monday 31 August 2020,   14:15,
mathematical physics end-of-summer seminar day (Kytölä)

Aapo Pajala
Euclidean Distance Geometry and Its Applications in 3D Genome Reconstruction (Bachelor thesis presentation)
Monday 31 August 2020,   14:00,

Konstantin Izyurov
Asymptotics of determinants of discrete Laplacians
Monday 31 August 2020,   13:30,
mathematical physics end-of-summer seminar day (Kytölä)

Tuomas Tuukkanen
Probabilistic Liouville conformal field theory
Monday 31 August 2020,   11:45,
mathematical physics end-of-summer seminar day (Kytölä)

Osama Abuzaid
Monday 31 August 2020,   11:00,
mathematical physics end-of-summer seminar day (Kytölä)

Nerissa Shakespeare
Äärellisistä heijastusryhmistä (kandiesitelmä)
Monday 31 August 2020,   10:00,
matematiikan kandiseminaari

Perttu Saarela
Hansen polytopes of split graphs (Bachelor thesis presentation)
Friday 28 August 2020,   11:30,   U5 (U147)

Kai Hippi
Tutte-polynomien laskeminen (Kandiesitelmä)
Friday 28 August 2020,   10:45,   U5 (U147)

Meri Aho
Cyclic flats and graphical matroids (Bachelor thesis presentation)
Friday 28 August 2020,   10:00,   U5 (U147)

Selim Virtanen
Johdatus p-adisiin lukuihin ja lokaali-globaali-periaatteeseen (kandiesitelmä)
Thursday 27 August 2020,   14:00,   U5 and Zoom

Okko Makkonen
Tuplaavuus ja Sobolevin upotus (Kandityöesitelmä)
Thursday 27 August 2020,   13:15,   U5

Joonas Laaksonen
Legendren sarjakehitelmän suppenemisesta 2D:ssä (Kandiesitelmä)
Thursday 27 August 2020,   11:00,

Etna Lindy
Parameter estimation for group-based phylogenetic models in the probability and Fourier coordinates (Bachelor thesis presentation)
Wednesday 26 August 2020,   11:00,   zoom

Aleksi Lahti
Classifying critical points of the Euclidean Distance from real plane curves (Bachelor thesis presentation)
Wednesday 26 August 2020,   10:00,   zoom

Hide past events

Page content by: webmaster-math [at] list [dot] aalto [dot] fi