Browsing by Title

Aikasarjoissa ilmenevien ei-normaalisten piirteiden mallintamiseen voidaan käyttää epälineaarisia aikasarjamalleja, joista erityisesti tutkielmassa tarkastellaan autoregressiivisia sekoitusmalleja. Autoregressiiviset sekoitusmallit määritellään sekoituksena lineaarisista autoregressiivisista malleista ja erona eri sekoitusmallien välillä on niiden sekoitussuhteiden määrittely. Autoregressiivisella GMAR (Gaussian Mixture Autoregressive)-sekoitusmallilla on houkuttelevia teoreettisia ominaisuuksia, sillä sen stationaarinen jakauma tunnetaan ja sen stationaarisuusehto ja ergodisuus voidaan johtaa ilman lisärajoituksia parametreille. Kuitenkin sekoitussuhteiden monimutkaisesta määrittelystä johtuen sen parametrien estimointi käyttäen kirjallisuudessa usein käytettyä EM-algoritmia on hankalaa. Tästä syystä tutkielmassa selvitetään mahdollisuutta käyttää parametrien estimoinnissa kaksivaiheista menetelmää, jossa geneettisen algoritmin avulla etsitään alkuarvoja gradienttiperusteiselle optimointialgoritmille. Parametrien estimoinnin lisäksi tutkielmassa tarkastellaan mallinvalintaa osana estimointiprosessia. Tarkasteltavia työkaluja sopivan mallin etsinnässä ovat informaatiokriteerit sekä erilaiset kvantiiliresiduaaleihin perustuvat testit, joiden avulla voidaan tehdä mallidiagnostiikkaa tavallisten residuaalien tapaan myös silloin, kun tavallisia residuaaleja ei voida käyttää. Lisäksi tarkastellaan ennusteiden laskemista simulaatioiden avulla ja esitetään miten GMAR-mallia voidaan simuloida. Tutkielman empiirisessä osassa tarkastellaan kahta esimerkkiä, joista ensimmäisessä keskitytään estimointiin, mallinvalitaan ja diagnostiikkaan. Tässä esimerkissä aineistona käytetään yhdysvaltain kuukausittaista inflaatiota vuodesta 1975 vuoteen 2015. Toisessa empiirisessä esimerkissä tarkastellaan tuulen nopeuksia päivittäisen aineiston avulla ja keskitytään erityisesti ennusteiden laskemiseen. Tuulen nopeutta mittaava aineisto on ei-negatiivinen aikasarja ja siksi esimerkissä tarkastellaan estimointia logaritmoidun sarjan avulla ja alkuperäisen sarjan ennustamista. Tutkielman tulosten perusteella kaksivaiheinen estimointi käyttäen geneettistä algoritmia toimii GMAR-mallin tapauksessa hyvin ja kohtuullisessa ajassa.

Tämä Pro Gradu -tutkielma käsittelee työntekijän eläkelain (TyEL) nykyisin käytössä olevan Gompertz-kuolevuusmallin sopivuutta kuvaamaan TyEL:n selektiä kuolevuutta vanhuuseläkeliikkeessä. Tutkielman tarkoitus on nostaa esiin nykymallin ongelmakohdat ja esitellä Gompertz-kuolevuusmallin laajennus, joka ainakin osin korjaisi nykymallin puutteita pysymällä kuitenkin järjestelmätekniseltä kannalta tarpeeksi yksinkertaisena. Keskeisiksi teemoiksi tutkielmassa nousee elämänvaravakuutuksen yleisen teorian sekä työntekijän eläkelain vanhuuseläkeliikkeen vakuutustekniikan lisäksi toteutuvan kuolevuuden ennustaminen Lee-Miller-mallin sovelluksella ja laajennetun Gompertz-kuolevuusmallin parametrien estimointi havaintoaineistosta. Havaintoaineistona tutkielmassa käytetään sekä Tilastokeskuksen väestökuolevuuksia että TyEL:n riskiperusteanalyysien mukaisia toteutuneita rahapainotettuja kuolevuuksia. Ennustemalli on laadittu käyttäen R- ja Excel-ohjelmistoja (ml. VBA). Kuolevuusmallin laajennusta on kehitetty työeläkejärjestelmässä vuodesta 2014 laskuperusteasiain neuvottelukunnan alaisessa kuolevuusperustejaoksessa, jonka sihteerinä tutkielman kirjoittaja on toiminut vuosina 2014 ja 2015. Puheenjohtajayhtiönä kyseisinä vuosina on toiminut Keskinäinen Työeläkevakuutusyhtiö Elo ja puheenjohtajana tämän tutkielman toinen tarkastaja Tuomas Hakkarainen. Kuolevuusperustejaoksessa on edustus jokaisen työeläkeyhtiön lisäksi eläkekassoilla ja eläkesäätiöillä, Sosiaali- ja terveysministeriöllä, Eläketurvakeskuksella sekä Kevalla (julkisen puolen eläkkeet). Kuolevuusmallin valinnalla ja osuvuudella on merkitystä vanhuuseläkeliikkeessä, sillä se määrää pääoma-arvokertoimet, joilla varaudutaan vastaisten ja alkaneiden vanhuuseläkkeiden suorituksiin tulevaisuudessa. Tutkielmassa esitelty uusi kuolevuusmalli otetaan käyttöön vuoden 2017 eläkeuudistuksen yhteydessä, eli ensimmäisen kerran vanhuuseläkeliikkeen vanhuuseläkevastuut lasketaan sen mukaisina vuoden 2016 lopussa. Vanhuuseläkemaksu määräytyy uuden mallin mukaisesti vuodesta 2017 alkaen.

This thesis examines the implementation of general purpose graphics processing unit (GPGPU) acceleration to a non-equilibrium Green’s function (NEGF) equation solver in the context of a computational photoelectrochemical (PEC) cell model. The goal is to find out whether GPGPU acceleration of the NEGF equation solver is a viable option. The current model does not yet have electron-photon scattering, but from the results it is possible to assess the viability of GPGPU acceleration in the case of a complete PEC cell model. The viability of GPGPU acceleration was studied by comparing the performance difference of two graphics processing unit (GPU) solutions against a multi core central processing unit (CPU) solution. The difference between the two GPU solutions was in the used floating-point precision. The GPU solutions would use LU factorization to solve the NEGF equations, and the CPU solution a banded solver (Gauss tridiagonal) provided by Scipy Python package. The performance comparison was done on multiple different GPU and CPU hardware. The electrical transport properties of the PEC cell were modeled by a self-consistent process in which the NEGF and Poisson equations were solved iteratively. The PEC cell was described as a semiconductor device connected with a metal and electrolyte contacts. The device was assumed to be a simple one dimensional tight-binding atom chain made of GaAs, where the transverse modes in the y–z plane are treated with a logarithmic function. The computational model did lack electron-photon scattering, which would be implemented in the future. From the benchmark results, it can be concluded that the GPGPU acceleration via LU factorization is not a viable option in the current code or in the complete model with electron-photon scattering and the assumed approximations. The parallel multi-core CPU code generally outperformed the GPU codes. The key weakness of the GPU code was the usage of LU factorization. Despite of this, there could be an opportunity for GPGPU acceleration if a more complex lattice structure and more exact scattering terms would be used. Also, a GPU accelerated tridiagonal solver could be a possible solution.

Learning a model over possible actions and using the learned model to maximize the obtained reward is an integral part of many applications. Trying to simultaneously learn the model by exploring state space and maximize the obtained reward using the learned model is an exploitation-exploitation tradeoff. Gaussian process upper confidence bound (GB-UCB) algorithm is an effective method for balancing between exploitation and exploration when exploring spatially dependent data in n-dimensional space. The balance between exploration and exploitation is required to limit the amount of user feedback required to achieve good prediction result in our context-based image retrieval system. The system starts with high amount of exploration and — as the confidence in the model increases — it starts exploiting the gathered information to direct the search towards better results. While the implementation of the GP-UCB is quite straightforward, it has time complexity of O(n^3) which limits its use in near real-time applications. In this thesis I present our reinforcement learning image retrieval system based on GP-UCB, with the focus on speed requirements for interactive applications. I also show simple methods to speed up the algorithm running time by doing some of the Gaussian process calculations on the GPU.

Alignment in genomics is the process of finding the positions where DNA strings fit best with one another, that is, where there are the least differences if they were placed side by side. This process, however, remains very computationally intensive, even with more recent algorithmic advancements in the field. Pseudoalignment is emerging as a new method over full alignment as an inexpensive alternative, both in terms of memory needed as well as in terms of power consumption. The process is to instead check for the existence of substrings within the target DNA, and this has been shown to produce good results for a lot of use cases. New methods for pseudoalignment are still evolving, and the goal of this thesis is to provide an implementation that massively parallelises the current state of the art, Themisto, by using all resources available. The most intensive parts of the pipeline are put on the GPU. Meanwhile, the components which run on the CPU are heavily parallelised. Reading and writing of the files is also done in parallel, so that parallel I/O can also be taken advantage of. Results on the Mahti supercomputer, using an NVIDIA A100, shows a 10 times end-to-end querying speedup over the best run of Themisto, using half the CPU cores as Themisto, on the dataset used in this thesis.

Due to the unique properties of foams, they can be found in many different applications in a wide variety of fields. The study of foams is also useful for the many properties they share with other phenomena, like impurities in cooling metals, where the impurities coarsen similarly to bubbles in foams. For these and other reasons foams have been studied extensively for over a hundred years and continue being an interesting area of study today due to new insights in both experimental and theoretical work and new applications waiting to be used and realized in different industries. The most impactful early work in the study of the properties of foams was done in the late 1800s by Plateau. His work was extended in the early to mid-1900s by Lifshitz, Slyozov, Wagner and von Neumann and by many more authors in recent years. The early work was mostly experimental or theoretical in the sense of performing mathematical calculations on paper, while the modern methods of study have kept the experimental part -- with more refined methods of measurement of course -- but shifted towards the implementation of the theory as simulations instead of solving problems on paper. In the early 90s Durian proposed a new method for simulating the mechanics of wet foams, based on repulsive spring-like forces between neighboring bubbles. This model was later extended to allow for the coarsening of the foam, and a slightly changed version of this model has been implemented in the code presented in this thesis. As foams consist of a very large number of bubbles, it is important to be able to simulate sufficiently large systems to realistically study the physics of foams. Very large systems have traditionally been too slow to simulate on the individual bubble level in the past, but thanks to the popularity of computer games and the continuous demand for better graphics in games, the graphics processing units have become very powerful and can nowadays be used to do highly parallel general computing. In this thesis, a modified version of Durian's wet foam model that runs on the GPU is presented. The code has been implemented in modern C++ using Nvidia's CUDA on the GPU. Using this program first a typical two-dimensional foam is simulated with 100000 bubbles. It is found that the simulation code replicates the expected behaviour for this kind of foam. After this, a more detailed analysis is done of a novel phenomenon of the separation of liquid and gas phases in low gas fraction foams that arises only with sufficiently large system sizes. It is found that the phase separation causes the foam to evolve as would a foam of higher gas fraction until the phases have mixed back together. It is hypothesized that the reason causing the phase separation is related to uneven energy distribution in the foam, which itself is related to jamming and uneven distribution of the sizes of the bubbles in the foam.

Grafeeni ja grafeenioksidi ovat kaksiulotteisia hiilimateriaaleja. Grafeeni koostuu pääosin sp2-hybridisoituneista hiiliatomeista. Grafeenioksidissa sp2-hybridisoituneiden hiiliatomien verkkoon on kiinnittynyt happea sisältäviä funktionaalisia ryhmiä ja osa hiiliatomeista on siten sp3-hybridisoituneita. Grafeenia voidaan erotella grafiitista erilaisin menetelmin tai valmistaa atomi kerrallaan. Grafeenioksidin valmistuksessa grafiitti hapetetaan esimerkiksi kloraateilla. Sekä grafeenia että grafeenioksidia voidaan muokata kovalenttisesti tai supramolekulaarisesti. Muokkauksella tavoitellaan usein materiaalin parempaa käsiteltävyyttä tai uudenlaisen toiminnallisuuden, kuten katalyyttisen aktiivisuuden tai sensoriominaisuuksien, saavuttamista. Katalyyttisissä prosesseissa muokkaamattoman grafeenin ja grafiitin sovelluksia ei ole monia. Sen sijaan grafeenioksidia on käytetty happokatalyytin tavoin sekä hapettimena monissa kemiallisissa prosesseissa, kuten alkyynien hydrauksissa, bentsyylialkoholien hapetuksessa aldehydiksi ja amiinien hapettavassa kytkennässä. Aktiivihiili on kolmiulotteinen huokoinen grafiitin kaltainen materiaali. Siitä on mahdollista valmistaa eri tavoin hapettuneita aktiivihiiliä typpihapon tai kuningasveden avulla. Hapetusolosuhteet vaikuttavat paljon syntyvän tuotteen funktionaalisten ryhmien jakaumaan. Eri tavoin valmistetuilla aktiivihiilillä on myös toisistaan poikkeava katalyyttinen aktiivisuus 2-aryyli-indolien homokytkentäreaktiossa.

Visualization is an indispensable method in the exploration of genomic data. However, the current state of the art in genome browsers – a class of interactive visualization tools – limit the exploration by coupling the visual representations with specific file formats. Because the tools do not support the exploration of the visualization design space, they are difficult to adapt to atypical data. Moreover, although the tools provide interactivity, the implementations are often rudimentary, encumbering the exploration of the data. This thesis introduces GenomeSpy, an interactive genome visualization tool that improves upon the current state of the art by providing better support for exploration. The tool uses a visualization grammar that allows for implementing novel visualization designs, which can display the underlying data more effectively. Moreover, the tool implements GPU-accelerated interactions that better support navigation in the genomic space. For instance, smoothly animated transitions between loci or sample sets improve the perception of causality and help the users stay in the flow of exploration. The expressivity of the visualization grammar and the benefit of fluid interactions are validated with two case studies. The case studies demonstrate visualization of high-grade serous ovarian cancer data at different analysis phases. First, GenomeSpy is being used to create a tool for scrutinizing raw copy-number variation data along with segmentation results. Second, the segmentations along with point mutations are used in a GenomeSpy-based multi-sample visualization that allows for exploring and comparing both multiple data dimensions and samples at the same time. Although the focus has been on cancer research, the tool could be applied to other domains as well.

Node classification is an important problem on networks in many different contexts. Optimizing the graph embedding has great potential to help improve the classification accuracy. The purpose of this thesis is to explore how graph embeddings can be exploited in the node classification task in the context of citation networks. More specifically, this thesis looks into the impact of different kinds of embeddings on the node classification, comparing their performance. Using three different similarity functions and different dimensions for the embedding vector ranging from 1 to 800, we examined the impact of graph embeddings on accuracy in node classification using three benchmark datasets: Cora, Citeseer, and PubMed. Our experimental results indicate that there are some common tendencies in the way dimensionality impacts the graph embedding quality regardless of the graph. We also established that some network-specific hyperparameter tuning clearly affects classification accuracy.

Introduction: The issue of climate change has emerged as a global challenge in response to the increasing consumption of natural resources. As the Information Technology (IT) sector has undergone significant growth in recent years, the implementation of environmentally sustainable practices which lower the environmental impact of software, such as electricity usage, has become imperative. The concept of green in software engineering seeks to address these challenges in the software engineering process. Methods: As the goal is to explore and evaluate different approaches to environmental sustainability in green in software engineering whilst also taking a look into the maturity and evidence level of research about the subject, this study adopts a systematic literature review approach. The search strings, search process and other relevant information are meticulously documented and explored in each step of the research process. Results: Green in software engineering has been identified as a promising field of research, but the absence of agreed-upon definitions and terminology often leads to research efforts replicating the previous studies without a clear reason as to why. The goal of increasing environmental sustainability is commonly agreed on in software engineering, but the concrete steps to achieve it are currently missing. Building a strong body of knowledge, common measurements and tooling to support them and increasing the knowledge about sustainability in the field of software engineering should all be taken into account in an effort to reach the environmental sustainability goals of tomorrow.