|
|
Possibilities of Neoclassical Geopolitics: Systemic Approach
Kofroň, Jan
The submitted dissertation tries to introduce neoclassical geopolitics as a viable approach to the study of international politics from geographical perspective. The dissertation is a compact of six already published articles and a common introduction highlighting main points of the articles and further discussing some issues which were (i) eliminated due to space constrains, or (ii) their significance is rather contextual, in the sense that they set the articles into broader discussions. The first part of the thesis (supported by two articles) deals with a current stage of political geography and geopolitics. The main result is that geopolitics is today a divided (sub)discipline, as geographers are mainly engaged in critical geopolitics and scholars of the International Relations continue in classical geopolitical reasoning (namely those who subscribe themselves under the label of neorealism). The main difference is that geographers consider space as an inter-subjective entity - socially constructed, whereas IR scholars tend to see space as an objective factor. In the combination with other epistemological differences, this different consideration of space has produced an exorbitant barrier between these two fields. One factor dividing the two approaches looms especially large - it is an arduous...
|
|
|
Analysis of Interface Automata with On-Demand Replication
Daniel, Jakub ; Kofroň, Jan (advisor)
Interface automaton is a model of software component behaviour based on finite state machines. It describes component's provided interface, the supported usage, and required interface, the usage of other components. A considerable number of components can be used in parallel with no bound on the level of parallelism. It is not necessary for the model to attempt to capture such unboundedness. An alternative approach is to allow an increment of the level of parallelism on- demand. This thesis analyses on a theoretical level and proposes a final form of an operation to perform such replication to allow creation of models of an arbitrary level of parallelism of certain parts of its behaviour.
|
|
| |
|
|
Boolean techniques in Knowledge representation
Chromý, Miloš ; Čepek, Ondřej (advisor) ; Mengel, Stefan (referee) ; Kofroň, Jan (referee)
Title: Boolean techniques in Knowledge representation Author: Miloš Chromý Department: Department of Theoretical Computer Science and Mathematical Logic Supervisor: Doc. RNDr. Ondřej Čepek, Ph.D., Department of Theoretical Com- puter Science and Mathematical Logic Abstract: In this thesis we will investigate switch-list representations of Boolean function and we will explore the biclique satisfiable formulas. Given a truth table representation of a Boolean function f the switch-list rep- resentation of f is a list of Boolean vectors from the truth table which have a different function value than the preceding Boolean vector in the truth table. We include this type of representation in the Knowledge Compilation Map [Dar- wiche and Marquis, 2002] and argue that switch-lists may in certain situations constitute a reasonable choice for a target language in knowledge compilation. First, we compare switch-list representations with a number of standard repre- sentations (such as CNF, DNF, and OBDD) with respect to their relative suc- cinctness. As a by-product of this analysis we also give a short proof of a long standing open question from [Darwiche and Marquis, 2002], namely the incom- parability of MODS (models) and PI (prime implicates) representations. Next, using the succinctness result between...
|
|
|
Forced Alignment via Neural Networks
Beňovič, Marek ; Kofroň, Jan (advisor) ; Hnětynka, Petr (referee)
Watching videos with subtitles in the original language is one of the most effective ways of learning a foreign language. Highlighting words at the moment they are pronounced helps to synchronize visual and auditory perception and increases learning efficiency. The method for aligning orthographic transcriptions to audio recordings is known as forced alignment. This work implements a tool for aligning transcript of YouTube videos with the speech in their audio recording, providing a web user interface with video player presenting the results. It integrates two state-of-the-art forced aligners based on Kaldi, first using standard HMM approach, second based on neural networks and compares their accuracy. Integrated aligners also provide a phone level alignment, which can be used for training statistical models in further speech recognition research. Work describes implementation and architectural concepts the tool is based on, which can be used in various software projects. 1
|
|
|
Central Asia and the New Silk Road: China and Russia - Cooperation or confrontation?
Kartsydze, Ketevan ; Romancov, Michael (advisor) ; Kofroň, Jan (referee)
Power relations are always a current and interesting topic for researchers. The intersection of economics and geopolitics today can give us a new perspective on traditional things. With addressing power relations in Central Asia, I have sought to contribute to the academic debate on the transfer of power from a different perspective. Power relations are always a current and interesting topic for researchers. Russia and China are two powerful states whose interests and goals intersect in Central Asia. China's economic expansion in the region leads to the question of whether or not the two power states can work together within one region. In this work, we will first look at the roles of Russia and China in Central Asia. In particular, will be analyzed economic and strategic cooperation with the help of regional organizations. this master thesis focuses on the period from 2001 to 2019. In the second part of the work by applying of power transition theory from American political scientist Abram Fim Kenneth Organski, will be analyzed whether the Russia-China relationship is closer to cooperation or confrontation.
|
|
|
STP solver for OpenSMT
Luňák, Václav ; Kofroň, Jan (advisor) ; Kučera, Petr (referee)
The Simple Temporal Problem is one of the fundamental scheduling problems. In the context of formal verification, it is closely related to SMT, where we can encounter it while solving the theory of difference logic. In this work we create a solver for difference logic as a part of the OpenSMT solver. We look at existing approaches and evaluate their applicability to OpenSMT. Then we analyze an algorithm based on exhaustive theory propagation and use it to create an efficient implementation of the solver. This implementation is tested and compared to other current SMT solvers, proving its comparable efficiency. 1
|
|
|
Prediction of results of tenis match in progress using machine learning
Konečný, Tomáš ; Kofroň, Jan (advisor) ; Pilát, Martin (referee)
This thesis focuses on applications of machine learning for prediction of a tennis match in progress. We propose three feedforward neural networks and compare them with a baseline model which assumes that individual points are independent and identically distributed. As a part of the comparison, odds from the betting exchange Betfair are used to estimate probability of winning. Features representing current match state and its past progress are proposed to make greater flexibility possible. Differences between models are inspected and an attempt to interpret them is made. 1
|
|
|
Anomaly detection for stock market trading data
Fusková, Martina ; Kofroň, Jan (advisor) ; Kliber, Filip (referee)
Stock trading is a very complex topic that involves a lot of challenging problems. One of these problems is anomaly detection in trading flow. Real-time anomaly detection in time series is a very complicated task and thus this issue is still open. The aim of this thesis is to research various models and algorithms that can be used for this task and try to find the most fitting ones. We develop models that detect anomalies based on the density properties of the data as well as statistical models and neural networks that detect anomalies based on the comparison of predicted data and actual data. As a result we propose models that can be further researched and used in real-time environment.
|
|
|
Does the size of an economy matter (a lot)? The economic potential of a country as a necessary or sufficient condition for regional dominance.
Boukal, Tomáš ; Kofroň, Jan (advisor) ; Karlas, Jan (referee)
The current economic rise of China entails one question. What significance can we attribute to the size of economy of countries for the accomplishment of regional hegemony? The ongoing debate agrees that economic might is a significant indicator of national power and thus the countries aspiring to dominate their regions should be displayed by lead in the economic sphere over other regional actors. Nevertheless, the consensus on the extent of this relationship has not yet been established. This thesis aims to contribute to the debate by determining whether economic potential constructed on GDP numbers is a necessary condition, a sufficient one, or neither of those, for regional dominance of countries. Based on the examination of historical cases from 1700 to 2010 and regions of Europe, Northeast Asia, and Western Hemisphere, we present a new indicator of regional dominance consisting of six sequential levels. By employing the "Necessary Condition Analysis", we argue that economic potential is a necessary but not a sufficient condition for higher levels of regional dominance with the effect of necessary condition denoted as "medium". Further, we provide a robustness check for the computation of economic potential using the Beckley's "GDP * GDP per capita" measure and Paul Bairoch's "Total industrial...
|
guest ::
RSS feed.