Browsing by master's degree program "Master's Programme in Mathematics and Statistics"

Tutkielmassa käsitellään avaruuden $\cc^n$ aitoja holomorfisia kuvauksia. Niiden määritelmät perustuvat aidon kuvauksen lauseeseen ja kuvausten holomorfisuuteen. Olkoon $\Omega,D\subset\cc^n$, ja olkoon $n>1$. Kuvaus $F:\Omega\to D$ on aito kuvaus, jos $F^{-1}(K)$ on kompakti $\Omega$:n osajoukko jokaiselle kompaktille joukolle $K\subset D$. Holomorfisuus tarkoittaa kuvauksen kompleksista analyyttisyyttä, kompleksista differentioituvuutta sekä sitä, että kuvaus toteuttaa Cauchy-Riemannin yhtälöt. Funktio $f$ on holomorfinen avaruuden $\cc^n$ avoimessa joukossa $\Omega$, jos sille pätee $f:\Omega\to\cc$, $f\in C^1(\Omega)$, ja jos se toteuttaa Cauchy-Riemannin yhtälöt $\overline{\partial}_jf=\frac{\partial f}{\partial\overline{z_j}}=0$ jokaiselle $j=1,\ldots,n$. Kuvaus $F=(f_1,\ldots,f_m):\Omega\to\cc^m$ on holomorfinen joukossa $\Omega$, jos funktiot $f_k$ ovat holomorfisia jokaisella $k=1,\ldots,m$. Jos $\Omega$ ja $D$ ovat kompleksisia joukkoja, ja jos $F:\Omega\to D$ on aito holomorfinen kuvaus, tällöin $F^{-1}(y_0)$ on joukon $\Omega$ kompakti analyyttinen alivaristo jokaiselle pisteelle $y_0\in D$. Aito kuvaus voidaan määritellä myös seuraavasti: Kuvaus $F:\Omega\to D$ on aito jos ja vain jos $F$ kuvaa reunan $\partial\Omega$ reunalle $\partial D$ seuraavalla tavalla: \[\text{jos}\,\{z_j\}\subset\Omega\quad\text{on jono, jolle}\,\lim_{j\to\infty}d(z_j,\partial\Omega)=0,\,\text{niin}\,\lim_{j\to\infty}d(F(z_j),\partial D)=0.\] Tämän määritelmän perusteella kuvausten $F:\Omega\to D$ tutkiminen johtaa geometriseen funktioteoriaan kuvauksista, jotka kuvaavat joukon $\partial\Omega$ joukolle $\partial D.$ Käy ilmi, että aidot holomorfiset kuvaukset laajenevat jatkuvasti määrittelyalueittensa reunoille. Holomorfisten kuvausten tutkiminen liittyy osaltaan Dirichlet-ongelmien ratkaisemiseen. Klassisessa Dirichlet-ongelmassa etsitään joukon $\partial\Omega\subset\mathbf{R}^m$ jatkuvalle funktiolle $f$ reaaliarvoista funktiota, joka on joukossa $\Omega$ harmoninen ja joukon $\Omega$ sulkeumassa $\overline{\Omega}$ jatkuva ja jonka rajoittuma joukon reunalle $\partial\Omega$ on kyseinen funktio $f$. Tutkielmassa käydään läpi määritelmiä ja käsitteitä, joista aidot holomorfiset kuvaukset muodostuvat, sekä avataan matemaattista struktuuria, joka on näiden käsitteiden taustalla. Tutkielmassa todistetaan aidolle holommorfiselle kuvaukselle $F:\Omega\to\Omega'$ ominaisuudet: $F$ on suljettu kuvaus, $F$ on avoin kuvaus, $F^{-1}(w)$ on äärellinen jokaiselle $w\in\Omega'$, on olemassa kokonaisluku $m$, jolle joukon $F^{-1}(w)$ pisetiden lukumäärä on $m$ jokaiselle $F$:n normaalille arvolle, joukon $F^{-1}(w)$ pisteiden lukumäärä on penempi kuin $m$ jokaiselle $F$:n kriittiselle arvolle, $F$:n kriittinen joukko on $\Omega'$:n nollavaristo, $F(V)$ on $\Omega'$:n alivaristo aina, kun $V$ on $\Omega$:n alivaristo, $F$ laajenee jatkuvaksi kuvaukseksi aidosti pseudokonveksien määrittelyjoukkojensa reunoille, $F$ kuvaa aidosti pseudokonveksin lähtöjoukkonsa jonon, joka suppenee epätangentiaalisesti kohti joukon reunaa, jonoksi joka suppenee hyväksyttävästi kohti kuvauksen maalijoukon reunaa, kuvaus $F$ avaruuden $\cc^n$ yksikköpallolta itselleen on automorfismi.

This paper contains Hilbert's Nullstellensatz, one of the fundamental theorems of algebraic geometry. The main approach uses the algebraic proof and all of the necessary background to understand and appreciate the Nullstellensatz. The paper also contains the Combinatorial Nullstellensatz, the proof and some applications. This thesis is aimed at any science student who wants to understand Hilbert's Nullstellensatz or gain intuitive insight. The paper starts with an explanation of the approach: the theorem is first presented in its algebraic form and the proof follows from the given background. After that, there are some more geometric definitions relating to common zero sets or varieties, and then Hilbert's Nullstellensatz is given in a slightly different manner. At the end of the paper some combinatorial definitions and theorems are introduced so that the Combinatorial Nullstellensatz can be proven. From this follows the Cauchy-Davenport theorem. At the very end of the introduction is presented a case of the theorem for the readers who do not yet know what Hilbert's Nullstellensatz is about. The second section of the paper contains all of the algebraic background material. It starts from the most basic definitions of groups and homomorphisms and then continues onto rings, ideals, radicals and quotient rings and relations between them. Some additional theorems are marked with a star to denote that the statement is not necessary for the algebraic proof of Hilbert's Nullstellensatz, but it might be necessary for the geometrical or combinatorial proofs. Since these statements are fully algebraic and mostly contain definitions that have been introduced early in the paper, they are placed in the algebraic section near similar theorems and definitions. The second section also contains fields, algebras, polynomials and transcendental elements. In Sections 3 and 4 we follow the algebraic proof of Daniel Allcock and expand on some steps. Section 3 contains Zariski's Lemma and a form of the Weak Nullstellensatz, along with their proofs. These statements allow us to skip Noetherian and Jacobson rings by finding the necessary conclusions from polynomial rings over fields. This section also includes an existence conclusion that also can be found in other literature under the name of Weak Nullstellensatz. Afterwards, Section 4 follows Allcock's proof of Hilbert's Nullstellensatz, working from the Weak Nullstellensatz and applying the Rabinowitsch trick. Section 5 explains how the Nullstellensatz brings together algebra and geometry. It is split up into three parts: first some preliminary definitions and theorems are defined. One of the fundamental definitions is a variety, which is simply the common zero set of some polynomials. After the new definitions we again meet Hilbert's Nullstellensatz and show that it is equivalent to the previous form. Using the newfound equivalence we show that varieties constitute a topology and consider the dimension of such spaces. Lastly, we show the equivalence of algebra homomorphisms and the corresponding variety morphisms. Section 6 slightly diverges and considers the Combinatorial Nullstellensatz introduced by Noga Alon. This theorem resembles Hilbert's Nullstellensatz, yet is different when looked at carefully. We consider the original proof by Alon, and additionally show how the Combinatorial Nullstellensatz follows from Hilbert's Nullstellensatz, using the paper by Goel, Patil and Verma. We conclude the paper by proving the Cauchy-Davenport theorem, using the Combinatorial Nullstellensatz. This paper does not delve deep in any particular topic. Quite the contrary, it connects many different theorems and shows equivalences between them while referencing additional reading material.

Online hypothesis testing occurs in many branches of science. Most notably it is of use when there are too many hypotheses to test with traditional multiple hypothesis testing or when the hypotheses are created one-by-one. When testing multiple hypotheses one-by-one, the order in which the hypotheses are tested often has great influence to the power of the procedure. In this thesis we investigate the applicability of reinforcement learning tools to solve the exploration – exploitation problem that often arises in online hypothesis testing. We show that a common reinforcement learning tool, Thompson sampling, can be used to gain a modest amount of power using a method for online hypothesis testing called alpha-investing. Finally we examine the size of this effect using both synthetic data and a practical case involving simulated data studying urban pollution. We found that, by choosing the order of tested hypothesis with Thompson sampling, the power of alpha investing is improved. The level of improvement depends on the assumptions that the experimenter is willing to make and their validity. In a practical situation the presented procedure rejected up to 6.8 percentage points more hypotheses than testing the hypotheses in a random order.

Electrical impedance tomography is a differential tomography method where current is injected into a domain and its interior distribution of electrical properties are inferred from measurements of electric potential around the boundary of the domain. Within the context of this imaging method the forward problem describes a situation where we are trying to deduce voltage measurements on a boundary of a domain given the conductivity distribution of the interior and current injected into the domain through the boundary. Traditionally the problem has been solved either analytically or by using numerical methods like the finite element method. Analytical solutions have the benefit that they are efficient, but at the same time have limited practical use as solutions exist only for a small number of idealized geometries. In contrast, while numerical methods provide a way to represent arbitrary geometries, they are computationally more demanding. Many proposed applications for electrical impedance tomography rely on the method's ability to construct images quickly which in turn requires efficient reconstruction algorithms. While existing methods can achieve near real time speeds, exploring and expanding ways of solving the problem even more efficiently, possibly overcoming weaknesses of previous methods, can allow for more practical uses for the method. Graph neural networks provide a computationally efficient way of approximating partial differential equations that is accurate, mesh invariant and can be applied to arbitrary geometries. Due to these properties neural network solutions show promise as alternative methods of solving problems related to electrical impedance tomography. In this thesis we discuss the mathematical foundation of graph neural network approximations of solutions to the electrical impedance tomography forward problem and demonstrate through experiments that these networks are indeed capable of such approximations. We also highlight some beneficial properties of graph neural network solutions as our network is able to converge to an arguably general solution with only a relatively small training data set. Using only 200 samples with constant conductivity distributions, the network is able to approximate voltage distributions of meshes with spherical inclusions.

In this thesis we will look at the asymptotic approach to modeling randomly weighted heavy-tailed random variables and their sums. The heavy-tailed distributions, named after the defining property of having more probability mass in the tail than any exponential distribution and thereby being heavy, are essentially a way to have a large tail risk present in a model in a realistic manner. The weighted sums of random variables are a versatile basic structure that can be adapted to model anything from claims over time to the returns of a portfolio, while giving the primary random variables heavy-tails is a great way to integrate extremal events into the models. The methodology introduced in this thesis offers an alternative to some of the prevailing and traditional approaches in risk modeling. Our main result that we will cover in detail, originates from "Randomly weighted sums of subexponential random variables" by Tang and Yuan (2014), it draws an asymptotic connection between the tails of randomly weighted heavy-tailed random variables and the tails of their sums, explicitly stating how the various tail probabilities relate to each other, in effect extending the idea that for the sums of heavy-tailed random variables large total claims originate from a single source instead of being accumulated from a bunch of smaller claims. A great merit of these results is how the random weights are allowed for the most part lack an upper bound, as well as, be arbitrarily dependent on each other. As for the applications we will first look at an explicit estimation method for computing extreme quantiles of a loss distributions yielding values for a common risk measure known as Value-at-Risk. The methodology used is something that can easily be adapted to a setting with similar preexisting knowledge, thereby demonstrating a straightforward way of applying the results. We then move on to examine the ruin problem of an insurance company, developing a setting and some conditions that can be imposed on the structures to permit an application of our main results to yield an asymptotic estimate for the ruin probability. Additionally, to be more realistic, we introduce the approach of crude asymptotics that requires little less to be known of the primary random variables, we formulate a result similar in fashion to our main result, and proceed to prove it.

Improving the quality of medical computed tomography reconstructions is an important research topic nowadays, when low-dose imaging is pursued to minimize the X-ray radiation afflicted on patents. Using lower radiation doses for imaging leads to noisier reconstructions, which then require postprocessing, such as denoising, in order to make the data up to par for diagnostic purposes. Reconstructing the data using iterative algorithms produces higher quality results, but they are computationally costly and not quite powerful enough to be used as such for medical analysis. Recent advances in deep learning have demonstrated the great potential of using convolutional neural networks in various image processing tasks. Performing image denoising with deep neural networks can produce high-quality and virtually noise-free predictions out of images originally corrupted with noise, in a computationally efficient manner. In this thesis, we survey the topics of computed tomography and deep learning for the purpose of applying a state-of-the-art convolutional neural network for denoising dental cone-beam computed tomography reconstruction images. We investigate how the denoising results of a deep neural network are affected if iteratively reconstructed images are used in training the network, as opposed to using traditionally reconstructed images. The results show that if the training data is reconstructed using iterative methods, it notably improves the denoising results of the network. Also, we believe these results can be further improved and extended beyond the case of cone-beam computed tomography and the field of medical imaging.

Työ käsittelee tunnettujen virtausdynamiikan yhtälöiden, Navier-Stokesin yhtälöiden ja Eulerin yhtälöiden välistä yhteyttä ja näiden ratkaisujen välisiä suppenemisehtoja. Työn ensimmäisessä kappaleessa esitellään työlle tärkeät perustiedot sisältäen esimerkiksi heikon derivaatan ja Sobolev-avaruuksien määritelmät ja useamman tärkeän funktioavaruuden ja jälkilauseiden määritelmät. Työn toinen kappale käsittelee tarkemmin Navier-Stokesin yhtälöitä ja Eulerin yhtälöitä. Kappaleessa esitellään ensin Navier-Stokesin yhtälöiden määritelmät ja sen jälkeen esitellään määritelmä ratkaisun olemassaololle. Kappaleen päätteeksi esitellään myös Eulerin yhtälön määritelmä. Neljännessä kappaleessa esitellään tutkielman pääaihe, eli Navier-Stokesin ja Eulerin yhtälöiden ratkaisujen välinen yhteys viskositeettitermin lähestyessä nollaa. Kappaleessa esitellään Tosio Katon tulos, jossa annetaan ekvivalentteja ehtoja sille, että Navier-Stokesin yhtälön heikko ratkaisu suppenee viskositeetin supetessa kohti Eulerin yhtälön ratkaisua. Tämä tulos todistetaan tutkiel- massa yksityiskohtaisesti. Lopuksi työn viimeisessä kappaleessa esitellään James. P. Kelliherin lisäykset Katon tuloksiin, jotka näyttävät, että Navier-Stokesin yhtälön ratkaisun u gradientti ∇u voidaan korvata ratkaisun u pyörteisyydellä ω(u). Kuten aiemmassa kappaleessa, niin myös tämä tulos esitellään yksityiskohtaisesti työssä. Työssä on vaadittu laajaa ymmärrystä monelta eri matematiikan osa-alueelta. Työn toinen kappale sisältää pitkälti analyysin metodeja sivuten muun muassa funktionaalianalyysiä ja funktioavaruuksien teoriaa. Kolmannessa ja neljännessä kappaleessa keskitytään pitkälti osittais- differentiaaliyhtälöiden teoriaan. Lisäksi työssä käsitellään laajalti myös reaalianalyysin aiheita. Päälähteinä työssä on käytetty Lawrence C. Evansin ”Partial Differential Equations” -teosta, Tosio Katon artikkelia ”Remarks on Zero Viscosity Limit” ja James P. Kelliherin artikkelia ”On Kato’s conditions for vanishing viscosity limit”.

Työssä todistetaan Delignen–Mumfordin kompaktifikaatiolause. Delignen–Mumfordin kompaktifikaatiolause sanoo, että tietyillä ehdolla jonolle saman signaturen hyperbolisia pintoja on olemassa rajapinta, josta on olemassa diffeomorfismit jokaiseen tämän jonon jäseneen ja että näillä diffeomorfismeilla nykäistyt metriikat suppenevat rajapinnalla rajametriikkaan. Työssä Delignen–Mumfordin kompaktifikaatiolause todistetaan osoittamalla vastaava tulos ensin hyperbolisten pintojen yksinkertaisille rakennuspalikoille. Työ todistaa vastaavan tuloksen ensin Y -palan kanonisille kauluksille ja käyttää tätä tulosta todistamaan vastaavan tuloksen Y -paloille. Tämän jälkeen työ todistaa jokaisen hyperbolisen pinnan olevan rakennettavissa Y -paloista antaen välittömästi tuloksen kaikille hyperbolisille pinnoille.

Bonus-malus systems are used globally to determine insurance premiums of motor liability policy-holders by observing past accident behavior. In these systems, policy-holders move between classes that represent different premiums. The number of accidents is used as an indicator of driving skills or risk. The aim of bonus-malus systems is to assign premiums that correspond to risks by increasing premiums of policy-holders that have reported accidents and awarding discounts to those who have not. Many types of bonus-malus systems are used and there is no consensus about what the optimal system looks like. Different tools can be utilized to measure the optimality, which is defined differently according to each tool. The purpose of this thesis is to examine one of these tools, elasticity. Elasticity aims to evaluate how well a given bonus-malus system achieves its goal of assigning premiums fairly according to the policy-holders’ risks by measuring the response of the premiums to changes in the number of accidents. Bonus-malus systems can be mathematically modeled using stochastic processes called Markov chains, and accident behavior can be modeled using Poisson distributions. These two concepts of probability theory and their properties are introduced and applied to bonus-malus systems in the beginning of this thesis. Two types of elasticities are then discussed. Asymptotic elasticity is defined using Markov chain properties, while transient elasticity is based on a concept called the discounted expectation of payments. It is shown how elasticity can be interpreted as a measure of optimality. We will observe that it is typically impossible to have an optimal bonus-malus system for all policy-holders when optimality is measured using elasticity. Some policy-holders will inevitably subsidize other policy-holders by paying premiums that are unfairly large. More specifically, it will be shown that, for bonus-malus systems with certain elasticity values, lower-risk policy-holders will subsidize the higher-risk ones. Lastly, a method is devised to calculate the elasticity of a given bonus-malus system using programming language R. This method is then used to find the elasticities of five Finnish bonus-malus systems in order to evaluate and compare them.

The presence of 1/f type noise in a variety of natural processes and human cognition is a well-established fact, and methods of analysing it are many. Fractal analysis of time series data has long been subject to limitations due to the inaccuracy of results for small datasets and finite data. The development of artificial intelligence and machine learning algorithms over the recent years have opened the door to modeling and forecasting such phenomena as well which we do not yet have a complete understanding of. In this thesis principal component analysis is used to detect 1/f noise patterns in human-played drum beats typical to a style of playing. In the future, this type of analysis could be used to construct drum machines that mimic the fluctuations in timing associated with a certain characteristic in human-played music such as genre, era, or musician. In this study the link between 1/f-noisy patterns of fluctuations in timing and the technical skill level of the musician is researched. Samples of isolated drum tracks are collected and split into two groups representing either low or high level of technical skill. Time series vectors are then constructed by hand to depict the actual timing of the human-played beats. Difference vectors are then created for analysis by using the least-squares method to find the corresponding "perfect" beat and subtracting them from the collected data. These resulting data illustrate the deviation of the actual playing from the beat according to a metronome. A principal component analysis algorithm is then run on the power spectra of the difference vectors to detect points of correlation within different subsets of the data, with the focus being on the two groups mentioned earlier. Finally, we attempt to fit a 1/f noise model to the principal component scores of the power spectra. The results of the study support our hypothesis but their interpretation on this scale appears subjective. We find that the principal component of the power spectra of the more skilled musicians' samples can be approximated by the function $S=1/f^{\alpha}$ with $\alpha\in(0,2)$, which is indicative of fractal noise. Although the less skilled group's samples do not appear to contain 1/f-noisy fluctuations, its subsets do quite consistently. The opposite is true for the first-mentioned dataset. All in all, we find that a much larger dataset is required to construct a reliable model of human error in recorded music, but with the small amount of data in this study we show that we can indeed detect and isolate defining rhythmic characteristics to a certain style of playing drums.