National Library of Finland DSpace Services
Not a member yet
86391 research outputs found
Sort by
Urvalsguide 2025 : studier för grundexamina i militärvetenskaper
No full textei tietoa saavutettavuudest
Ricoeur on Peace and Violence in the Ontology of Creation
No full textThis essay is the first of two seeking to draw out an ontology of peace from Paul Ricoeur’s thought. This first essay (Part I) argues that Ricoeur’s hermeneutics of creation provides the best starting point because of its insistence on the goodness of created being. Ricoeur develops this conviction from his reading of the biblical creation accounts, which I follow through three texts from three periods of Ricoeur’s work. In The Symbolism of Evil, Ricoeur show that peace rather than violence is most fundamental to creation. In his essay “On the Exegesis of Gen 1:1-2:4a,” he expands his interpretation to consider the combat imagery in the Psalms, but shows how the text interprets the separation and ordering of creation as a work of providential wisdom rather than violence. In Thinking Biblically, Ricoeur complicates his earlier hermeneutics of creation by bringing in themes of mastery, chaos, and fragility—three themes that need careful interpretation in order to preserve Ricoeur’s earlier emphasis on the goodness and peacefulness of creation. This preservation is possible, I argue, by recovering Ricoeur’s early Christological reflections in The Symbolism of Evil, which point to the hope of an ultimate, eschatological victory over violence. I conclude by arguing that Ricoeur’s hermeneutics can help us to imagine peace, which is crucial to the practice of peace
FaultLocalizer 2.0 Enhanced Fault Location for Python Programs Using Spectrum Analysis
No full textFault localization is a challenging and time-consuming aspect of software maintenance, often requiring substantial effort to identify the source of issues within complex software systems. To streamline this process, we introduce FaultLocalizer 2.0, a dedicated tool developed to aid developers in accurately identifying fault locations within code. FaultLocalizer 2.0 integrates spectrum-based fault localization with program slicing techniques to efficiently narrow down areas requiring attention, thus reducing debugging time. The tool generates detailed fault reports, which include a call graph for analyzing inter-procedural dependencies and annotated control flow graphs that mark areas needing further analysis for potential repairs. Built in Python, FaultLocalizer 2.0 offers an intuitive graphical user interface, enabling visualization of both call and control flow graphs derived from dynamic execution data. The tool takes as input the source code, a suite of passed and failed test cases, and program spectra data collected during test runs. Tested on a repository of Python programs with known faults, FaultLocalizer 2.0 demonstrated its effectiveness in significantly lowering debugging and repair efforts, with promising applications in fault localization for Python programs with complex branching structures
Evaluating Transformer Models For Document Question Answering : A Comparative Study On Multilingual And Normative Texts
No full textDocument Question Answering (DQA) systems have gained significant traction in recent years, especially with the rise of large language models (LLMs). However, the application of these models to normative documents, such as legal texts, guidelines, or emergency instructions, presents unique chal- lenges. Normative documents often contain formal, domain- specific language, making it difficult for DQA models to extract precise answers. This study evaluates the performance of various transformer-based models on a bilingual dataset of university normative documents, comprising English and Turk- ish texts. For English, models BERT, DistilBERT, ALBERT, RoBERTa, T5, and Longformer are analyzed. For Turkish, XLM-R, and fine-tuned Turkish-specific models BERTurk and PEFT mT0 are utilized to address the unique morphological challenges of the Turkish language. Evaluation metrics, in- cluding token-level accuracy, F1 score, and execution time, are used to analyze the models’ performance in extracting relevant information. This research identifies the most effective models for processing normative documents in each language and highlights the limitations of current DQA systems in handling complex, formal texts across linguistic boundaries
A systems-based approach towards analyzing the human heart
No full textThis study aims to develop a simulation solution (software package) that integrates mathematical, statistical, and signal processing techniques alongside machine learning methods for analyzing the human heart. This includes interactions between electrocardiogram (ECG), photoplethysmography (PPG), and electroencephalogram (EEG) signals, particularly during critical episodes such as cardiac arrest. These signals capture the heartbeat and the heart’s regulation by the autonomic nervous system (ANS). Analyzing these signals together provides deeper insights into the heart than when they are studied separately. Essentially, these signal types are valuable because they enable the heart to be assessed as part of a larger system.
The core of the solution is designed to analyze heartbeats from a single-lead, unprocessed signal. It uses this analysis to evaluate brain activity during identifiable events such as cardiac arrest. This is achieved by treating individual heartbeats (waveforms) as states within a simulation. The simulation manages the classification of these individual heartbeats (ECG) and uses this classification to perform event-matching with brain activity. This approach aligns with building a Kalman Filter geared toward detecting anomalies.
The sources used in this study include I-CARE and MIT-BIH Arrhythmia Databases. The I-CARE database is newly released in 2023. It offers synchronized ECG and EEG recordings, which makes it particularly valuable for assessing the heart as part of a larger physiological system. The MIT-BIH Arrhythmia Database is older, with its current release dating back to 2005 and the data originating from the 1970s. It has limitations in sample frequency range and data quality.
The main outcome demonstrates the capability to isolate specific heartbeats corresponding to distinct brain activity. The Positive Predictive Value (PPV) value for the convolutional neural network (CNN) using the MIT-BIH Arrhythmia Database is approximately 0.915, and the F1 score is around 0.87. The study also examines the connection between ECG and EEG through event-matching. The results indicate a clear difference in brain activity when the heart is pumping compared to when it is not. The peak amplitude alternates from -8.1 for prolonged cardiac arrest (x > 2s) to 13.2 microvolts (μV) for a normal heartbeat
Empirical Monte Carlo evidence on estimation of timing-of-events models
No full textThis article builds on the Empirical Monte Carlo simulation approach to study the estimation of Timing-of-Events (ToE) models. We exploit rich Swedish data of unemployed job seekers with information on participation in a training program to simulate placebo treatment durations. We first use these simulations to examine which covariates are key confounders to be included in dynamic selection models for training participation. The joint inclusion of specific short-term employment history indicators (notably, the share of time spent in employment), together with baseline socio-economic characteristics, regional and inflow timing information, is important to deal with selection bias. Next, we omit subsets of explanatory variables and estimate ToE models with discrete distributions for the ensuing systematic unobserved heterogeneity. In many cases, the ToE approach provides accurate effect estimates, especially if time-varying variation in the unemployment rate of the local labor market is taken into account. However, assuming too many or too few support points for unobserved heterogeneity may lead to large biases. Information criteria, in particular those penalizing parameter abundance, are useful to select the number of support points. A comparison with other duration models shows that a Stratified Cox model performs well with abundant multiple spells but less well when multiple spells are uncommon. The standard Cox regression model performs poorly in all configurations as it is unable to account for unobserved heterogeneity.peerReviewe
Strategizing the Shallows : Leveraging Multi-Agent Reinforcement Learning for Enhanced Tactical Decision-Making in Littoral Naval Warfare
No full textNaval engagements, though rare, present complex challenges for data-driven machine learning due to their intricate dynamics and the scarcity of empirical evidence. These conflicts are well-represented within the framework of Partially Observable Stochastic Games (POSG), which models the adversarial interactions between contending forces through decision-making agents, possible states, actions, observations, and probabilistic state transitions.
This research delineates the implementation of Multi-Agent Reinforcement Learning (MARL) algorithms, particularly Double Deep Q-Networks (DDQN) and Proximal Policy Optimization (PPO), in navigating the complexities inherent in strategic naval operations. Despite the operational challenges encountered, the findings of this study underscore the effectiveness of MARL in formulating and assessing tactical strategies. This contributes substantially to the enhancement of tactical planning and the introduction of novel strategic paradigms. Significantly, this investigation illuminates the transformative potential of MARL in naval strategy and decision-making processes, asserting its pivotal role in contemporary warfare analysis. This study not only confirms the applicability of MARL in complex scenarios but also highlights its capacity to revolutionize traditional approaches to military strategy
Kansainvälinen oikeus ja turvapaikan hakeminen Suomen itärajalla
No full textPyrkimys valvoa ja torjua ei-toivottua maahanmuuttoa on yleinen piirre valtioiden reaktioissa pakkomuuttoon. Suomen tämänhetkisessä tilanteessa ei ole mitään erityisen uutta tai ennen kokematonta – vastentahtoiset mahdolliset vastaanottajamaat ovat torjuneet ei -toivottuja pakkosiirtolaisia niin kauan kuin dokumentoitua muuttoliikettä on tapahtunut. (FitzGerald 2019) Viime vuosikymmeninä tästä vastentahtoisuudesta on tullut yhä kohdennetumpaa, laskelmoidumpaa, järjestelmällisempää ja julmempaa. Suomen reaktio noudattaa ennalta arvattavasti tätä kaavaa.ei tietoa saavutettavuudest
Sukella syvälle, sukella hiljaa : suomalaisen sukellusvenetaktiikan kehittyminen talvi- ja jatkosodassa
No full textSukellusvenelaivue ja siihen kuuluneet viisi sukellusvenettä olivat osa Merivoimien kokoonpanoa. Sukellusvenelaivue otti osaa sotatoimiin sekä talvi-, että jatkosodassa. Tässä tutkielmassa selvitetään, millaista taktiikkaa sukellusveneet käyttivät ja miten taktiikka kehittyi sotien aikana. Tähän tutkimuskysymykseen vastaamiseksi tutkielmassa tarkastellaan kolmea taktiikan kehitykselle tärkeimmäksi nostettua aihealuetta: alusten tekninen kunto ja suorituskyky, miehistön koulutus, sekä alusten käyttö ja tehtävät sotien aikana.
Alukset olivat talvisodan syttyessä jo koneistoiltaan kuluneita ja niitä kuvailtiinkin loppuun ajetuiksi. Tästä huolimatta ne palvelivat läpi sotavuosien ja niiden teknistä suorituskykyä paranneltiin jatkosodan loppuun saakka. Tutkielmassa selvitetään tämän teknisen kunnon ja sen ylläpidon vaikutusta sukellusvenetaktiikkaan. Teknisen kunnon lisäksi toinen kriittinen seikka menestyksekkäälle sukellusvenetaktiikalle oli osaava ja harjaantunut miehistö. Sukellusveneitä pidettiin erityisenä, vaativana aselajina ja koulutuksen taso oli korkea. Henkilöstö pyrittiin myös pitämään sukellusveneillä useiden vuosien ajan, parhaan mahdollisen kokemuksen ja osaamisen säilymiseksi laivueessa. Tekninen suorituskyky ja henkilöstön osaaminen loivat valmiudet toteuttaa sukellusveneille määriteltyjä tehtäviä. Näiden tehtävien tarkastelun kautta luodaan käsitys siitä, mihin sukellusveneet oikeasti sotatoimissa kykenivät ja kuinka paljon siihen vaikuttivat juuri edellä mainitut tekniset ja koulutukselliset seikat.
Tutkielma on toteutettu laadullisena tutkimuksena ja menetelmänä on aineistolähtöinen sisällönanalyysi. Lähdeaineiston rungon muodostavat Kansallisarkiston, sekä Merisotakoulun kirjaston arkiston aineistot, kuten Merivoimien operatiivisen osaston yleinen ja salainen kirjeenvaihto vuosilta 1939-1944, sekä erilaiset käsikirjoitukset 1930-luvulta. Toissijaisina lähteinä on käytetty aiempia kadetti- ja pro gradu - tutkielmia, sekä sukellusveneitä koskevaa tietokirjallisuutta.
Talvisodassa sukellusveneiden toiminta jäi verrattain vähäiseksi. Jatkosodan kahta ensimmäistä vuotta voidaan pitää laivueen menestyksekkäimpinä, niiden sisältäessä sekä useita laskettuja miinavaarannoksia, että onnistuneita vihollisalusten torpedointeja. Jatkosodan viimeiset vuodet taas laivue vietti hiljaiseloa, tehtävien määrän jäädessä vähäiseksi. Tutkielmassa selvitetään, miten alusten käyttämä taktiikka muuttui ja kehittyi näiden luonteeltaan hyvin erilaisten ajanjaksojen aikana. Muutoksen lisäksi tutkielmassa perehdytään tämän muutoksen taustalla olleisiin syihin, eli mitkä asiat vaikuttivat taktiikan kehityksee