Results 1-20 of 807.
((filter:inst)) AND ((type:#F03))

Bookmark and Share    
Full Text
See detailTransforming Time Series for Efficient and Accurate Classification
Li, Daoyuan UL

Doctoral thesis (2018)

Time series data refer to sequences of data that are ordered either temporally, spatially or in another defined order. They can be frequently found in a variety of domains, including financial data ... [more ▼]

Time series data refer to sequences of data that are ordered either temporally, spatially or in another defined order. They can be frequently found in a variety of domains, including financial data analysis, medical and health monitoring and industrial automation applications. Due to their abundance and wide application scenarios, there has been an increasing need for efficient machine learning algorithms to extract information and build knowledge from these data. One of the major tasks in time series mining is time series classification (TSC), which consists of applying a learning algorithm on labeled data to train a model that will then be used to predict the classes of samples from an unlabeled data set. Due to the sequential characteristic of time series data, state-of-the-art classification algorithms (such as SVM and Random Forest) that performs well for generic data are usually not suitable for TSC. In order to improve the performance of TSC tasks, this dissertation proposes different methods to transform time series data for a better feature extraction process as well as novel algorithms to achieve better classification performance in terms of computation efficiency and classification accuracy. In the first part of this dissertation, we conduct a large scale empirical study that takes advantage of discrete wavelet transform (DWT) for time series dimensionality reduction. We first transform real-valued time series data using different families of DWT. Then we apply dynamic time warping (DTW)-based 1NN classification on 39 datasets and find out that existing DWT-based lossy compression approaches can help to overcome the challenges of storage and computation time. Furthermore, we provide assurances to practitioners by empirically showing, with various datasets and with several DWT approaches, that TSC algorithms yield similar accuracy on both compressed (i.e., approximated) and raw time series data. We also show that, in some datasets, wavelets may actually help in reducing noisy variations which deteriorate the performance of TSC tasks. In a few cases, we note that the residual details/noises from compression are more useful for recognizing data patterns. In the second part, we propose a language model-based approach for TSC named Domain Series Corpus (DSCo), in order to take advantage of mature techniques from both time series mining and Natural Language Processing (NLP) communities. After transforming real-valued time series into texts using Symbolic Aggregate approXimation (SAX), we build per-class language models (unigrams and bigrams) from these symbolized text corpora. To classify unlabeled samples, we compute the fitness of each symbolized sample against all per-class models and choose the class represented by the model with the best fitness score. Through extensive experiments on an open dataset archive, we demonstrate that DSCo performs similarly to approaches working with original uncompressed numeric data. We further propose DSCo-NG to improve the computation efficiency and classification accuracy of DSCo. In contrast to DSCo where we try to find the best way to recursively segment time series, DSCo-NG breaks time series into smaller segments of the same size, this simplification also leads to simplified language model inference in the training phase and slightly higher classification accuracy. The third part of this dissertation presents a multiscale visibility graph representation for time series as well as feature extraction methods for TSC, so that both global and local features are fully extracted from time series data. Unlike traditional TSC approaches that seek to find global similarities in time series databases (e.g., 1NN-DTW) or methods specializing in locating local patterns/subsequences (e.g., shapelets), we extract solely statistical features from graphs that are generated from time series. Specifically, we augment time series by means of their multiscale approximations, which are further transformed into a set of visibility graphs. After extracting probability distributions of small motifs, density, assortativity, etc., these features are used for building highly accurate classification models using generic classifiers (e.g., Support Vector Machine and eXtreme Gradient Boosting). Based on extensive experiments on a large number of open datasets and comparison with five state-of-the-art TSC algorithms, our approach is shown to be both accurate and efficient: it is more accurate than Learning Shapelets and at the same time faster than Fast Shapelets. Finally, we list a few industrial applications that relevant to our research work, including Non-Intrusive Load Monitoring as well as anomaly detection and visualization by means for hierarchical clustering for time series data. In summary, this dissertation explores different possibilities to improve the efficiency and accuracy of TSC algorithms. To that end, we employ a range of techniques including wavelet transforms, symbolic approximations, language models and graph mining algorithms. We experiment and evaluate our approaches using publicly available time series datasets. Comparison with the state-of-the-art shows that the approaches developed in this dissertation perform well, and contribute to advance the field of TSC. [less ▲]

Detailed reference viewed: 36 (13 UL)
Full Text
See detailComputational Methods for Analysing Long-run Dynamics of Large Biological Networks
Yuan, Qixia UL

Doctoral thesis (2017)

Systems biology combines developments in the fields of computer science, mathematics, engineering, statistics, and biology to study biological networks from a holistic point of view in order to provide a ... [more ▼]

Systems biology combines developments in the fields of computer science, mathematics, engineering, statistics, and biology to study biological networks from a holistic point of view in order to provide a comprehensive, system level understanding of the underlying system. Recent developments in biological laboratory techniques have led to a slew of increasingly complex and large biological networks. This poses a challenge for formal representation and analysis of those large networks efficiently. To understand biology at the system level, the focus should be on understanding the structure and dynamics of cellular and organismal function, rather than on the characteristics of isolated parts of a cell or organism. One of the most important focuses is the long-run dynamics of a network, as they often correspond to the functional states, such as proliferation, apoptosis, and differentiation. In this thesis, we concentrate on how to analyse long-run dynamics of biological networks. In particular, we examine situations where the networks in question are very large. In the literature, quite a few mathematical models, such as ordinary differential equations, Petri nets, and Boolean networks (BNs), have been proposed for representing biological networks. These models provide different levels of details and have different advantages. Since we are interested in large networks and their long-run dynamics, we need to use ``coarse-grained" level models that focus on the system behaviour of the network while neglecting molecular details. In particular, we use probabilistic Boolean networks (PBNs) to describe biological networks. By focusing on the wiring of a network, a PBN not only simplifies the representation of the network, but it also captures the important characteristics of the dynamics of the network. Within the framework of PBNs, the analysis of long-run dynamics of a biological network can be performed with regard to two aspects. The first aspect lies in the identification of the so-called attractors of the constituent BNs of a PBN. An attractor of a BN is a set of states, inside which the network will stay forever once it goes in; thus capturing the network's long-term behaviour. A few methods have been discussed for computing attractors in the literature. For example, the binary decision diagram based approach and the satisfiability based approach. These methods, however, are either restricted by the network size, or can only be applied to synchronous networks where all the elements in the network are updated synchronously at each time step. To overcome these issues, we propose a decomposition-based method. The method works in three steps: we decompose a large network into small sub-networks, detect attractors in sub-networks, and recover the attractors of the original network using the attractors of the sub-networks. Our methods can be applied to both asynchronous networks, where only one element in the network is updated at each time step, and synchronous networks. Experimental results show that our proposed method is significantly faster than the state-of-the-art methods. The second aspect lies in the computation of steady-state probabilities of a PBN with perturbations. The perturbations of a PBN allow for a random, with a small probability, alteration of the current state of the PBN. In a PBN with perturbations, the long-run dynamics is characterised by the steady-state probability of being in a certain set of states. Various methods for computing steady-state probabilities can be applied to small networks. However, for large networks, the simulation-based statistical methods remain the only viable choice. A crucial issue for such methods is the efficiency. The long-run analysis of large networks requires the computation of steady-state probabilities to be finished as soon as possible. To reach this goal, we apply various techniques. First, we revive an efficient Monte Carlo simulation method called the two-state Markov chain approach for making the computations. We identify an initialisation problem, which may lead to biased results of this method, and propose several heuristics to avoid this problem. Secondly, we develop several techniques to speed up the simulation of PBNs. These techniques include the multiple central processing unit based parallelisation, the multiple graphic processing unit based parallelisation, and the structure-based parallelisation. Experimental results show that these techniques can lead to speedups from ten times to several hundreds of times. Lastly, we have implemented the above mentioned techniques for identification of attractors and the computation of steady-state probabilities in a tool called ASSA-PBN. A case-study for analysing an apoptosis network with this tool is provided. [less ▲]

Detailed reference viewed: 19 (11 UL)
Full Text
See detailEfficient and Secure Implementations of Lightweight Symmetric Cryptographic Primitives
Dinu, Dumitru Daniel UL

Doctoral thesis (2017)

This thesis is devoted to efficient and secure implementations of lightweight symmetric cryptographic primitives for resource-constrained devices such as wireless sensors and actuators that are typically ... [more ▼]

This thesis is devoted to efficient and secure implementations of lightweight symmetric cryptographic primitives for resource-constrained devices such as wireless sensors and actuators that are typically deployed in remote locations. In this setting, cryptographic algorithms must consume few computational resources and withstand a large variety of attacks, including side-channel attacks. The first part of this thesis is concerned with efficient software implementations of lightweight symmetric algorithms on 8, 16, and 32-bit microcontrollers. A first contribution of this part is the development of FELICS, an open-source benchmarking framework that facilitates the extraction of comparative performance figures from implementations of lightweight ciphers. Using FELICS, we conducted a fair evaluation of the implementation properties of 19 lightweight block ciphers in the context of two different usage scenarios, which are representatives for common security services in the Internet of Things (IoT). This study gives new insights into the link between the structure of a cryptographic algorithm and the performance it can achieve on embedded microcontrollers. Then, we present the SPARX family of lightweight ciphers and describe the impact of software efficiency in the process of shaping three instances of the family. Finally, we evaluate the cost of the main building blocks of symmetric algorithms to determine which are the most efficient ones. The contributions of this part are particularly valuable for designers of lightweight ciphers, software and security engineers, as well as standardization organizations. In the second part of this work, we focus on side-channel attacks that exploit the power consumption or the electromagnetic emanations of embedded devices executing unprotected implementations of lightweight algorithms. First, we evaluate different selection functions in the context of Correlation Power Analysis (CPA) to infer which operations are easy to attack. Second, we show that most implementations of the AES present in popular open-source cryptographic libraries are vulnerable to side-channel attacks such as CPA, even in a network protocol scenario where the attacker has limited control of the input. Moreover, we describe an optimal algorithm for recovery of the master key using CPA attacks. Third, we perform the first electromagnetic vulnerability analysis of Thread, a networking stack designed to facilitate secure communication between IoT devices. The third part of this thesis lies in the area of side-channel countermeasures against power and electromagnetic analysis attacks. We study efficient and secure expressions that compute simple bitwise functions on Boolean shares. To this end, we describe an algorithm for efficient search of expressions that have an optimal cost in number of elementary operations. Then, we introduce optimal expressions for first-order Boolean masking of bitwise AND and OR operations. Finally, we analyze the performance of three lightweight block ciphers protected using the optimal expressions. [less ▲]

Detailed reference viewed: 62 (24 UL)
See detailCONTRIBUTIONS TO THE STATISTICS OF RANDOM PROCESSES USING MALLIAVIN CALCULUS
Krein, Christian Yves Léopold UL

Doctoral thesis (2017)

In this dissertation we present several applications of Malliavin calculus, both to the statistical analysis of continuous time stochastic processes and to limit theorems for non-linear functionals of ... [more ▼]

In this dissertation we present several applications of Malliavin calculus, both to the statistical analysis of continuous time stochastic processes and to limit theorems for non-linear functionals of Gaussian Fields. Malliavin calculus extends techniques of classical calculus of variations from deterministic functions to random variables. In Malliavin calculus, the so called Malliavin derivative and its adjoint, the divergence operator, are combined with the theory of Hilbert spaces. Just as classical calculus, this theory has proved to be a powerful tool and its applications vary from the existence of densities, to the construction of estimators and the study of weak convergence of sequences of random variables and random vectors, with a special focus on normal approximations. The first part of the present document is essentially a generalization of a result of Privault and Réveillac (2008), which extends a seminal paper of Stein (1956). Stein has shown that, under certain conditions, there are biased estimators which perform better than the standard estimator for the mean of a multivariate normal vector. It has been shown by Privault and Réveillac that a similar statement holds for Gaussian processes and we shall present a generalization of their work to continuous time models, where the noise is either a chaotic Brownian martingale or a non-martingale noise living in the second Wiener chaos. This first part of the work corresponds to the paper "Drift estimation with non-gaussian noise using Malliavin Calculus" (2015) which has been published by the Electronic Journal of Statistics. In the second part of the work we give necessary and sufficient criteria for the convergence of sequences of random variables, living in a fixed sum of Wiener chaoses, to a limit which lives in the sum of the first two Wiener chaoses. Our results extend the important findings of Nualart and Peccati (2005), the so-called Fourth Moment Theorem, and a recent finding of Azmoodeh, Peccati and Poly (2014). Our criteria make use of the so-called Gamma-operators which are derived from scalar products of Malliavin derivatives and the infinitesimal generator of the Ornstein-Uhlenbeck semi-group, see for instance Azmoodeh, Peccati and Poly (2014). This part corresponds to the paper "Weak convergence on Wiener space: targeting the first two chaoses" (2017) which has been submitted to the Latin American Journal of Probability and Mathematical Statistics (ALEA). In the last part of the present work we consider a sequence living in a fixed Wiener chaos and converging in law to a normal variable. A second sequence is supposed to converge in law to a target variable which is the sum of a linear combination of independent chi-square distributed random variables and an independent normal variable. We derive conditions under which the sequence of random vectors, formed by both sequences of random variables, converges in law. We use again Gamma-operators and cumulants to derive necessary and sufficient conditions which can be seen as generalization of results of Peccati and Tudor (2005) for Gaussian limits in the case of sequences of random vectors which converge componentwise. We apply methods developed by Nourdin and Peccati (2009) to examine the rate of convergence of a sequence of double Wiener integrals towards a normal variable. [less ▲]

Detailed reference viewed: 21 (3 UL)
See detailLearning Finite Automata via Flexible State-Merging and Applications in Networking
Hammerschmidt, Christian UL

Doctoral thesis (2017)

Being able to model behavior described by a linear sequence of observations (such as log files) goes a long way towards better understanding the underlying processes. This improved understanding can be ... [more ▼]

Being able to model behavior described by a linear sequence of observations (such as log files) goes a long way towards better understanding the underlying processes. This improved understanding can be very helpful in a number of activities, ranging from software (reverse) engineering to network traffic analysis. The developments in this thesis were driven by specific goals in predicting (human) behaviors captured by a software appliance observing network traffic and user requests to specific resources. Its final contributions have exceeded the original goals of the project in two important ways: I present (1) a flexible learning algorithm for finite automata accompanied by theoretical underpinning and its implementation, a contribution towards better learning algorithms, and (2) applications of the algorithm to use-cases in computer networking and beyond. The central algorithm considered in the thesis is a blue-fringe state-merging automaton learning algorithm, conducting a greedy search over feasible solutions. Its key components are a heuristic to search for consistent merges and an evaluation metric to assess the quality of a merge by assigning scores to merges. I generalize this framework by making the heuristic components explicitly parametric. While state-merging algorithms were originally defined for probabilistic and non-probabilistic finite state machines and later used to derive algorithms for more extended models such as real-time automata, the work presented here extends the scope of the algorithms to a wide range of ad-hoc defined models as well as enables the user to implement modifications to the heuristic search process. These modifications help to account for domain knowledge and richer semantics of models with a regular language core. I provide an implementation and a Python interface of the flexible state-merging framework, including stream/online and interactive variants of the algorithm based on a C++ implementation of the blue-fringe greedy search algorithm called DFASAT. The algorithm and the framework encompass and improve upon state-of-the-art approaches. The application problems considered in this thesis can be seen as classical classification and anomaly detection tasks in machine learning. The application domain is network traffic analysis with a focus on network security. I discuss the problematic properties of data from computer networks and address how using automaton models can help mitigate them. I then use the flexible state-merging approach for host profiling. I show how to efficiently learn finite state automata as behavioral profiles. These profiles can serve as digital fingerprints and help to identify malicious traffic such as botnet traffic. Moreover, I show how communication profiles can be used for sequence clustering on NetFlow data to distinguish different behaviors over time. [less ▲]

Detailed reference viewed: 30 (9 UL)
Full Text
See detailAPPROACHES FOR IDENTIFICATION OF TRANSCRIPTIONAL AND POST-TRANSCRIPTIONAL REGULATORS OF MESENCHYMAL STEM CELL DIFFERENTIATION USING TIME-SERIES EPIGENOMIC DATA
Gerard, Déborah UL

Doctoral thesis (2017)

Gene regulatory networks (GRNs) control cellular differentiation and development and recapitulate the physical interactions between transcription factors (TFs) and their influence on their target genes ... [more ▼]

Gene regulatory networks (GRNs) control cellular differentiation and development and recapitulate the physical interactions between transcription factors (TFs) and their influence on their target genes that ultimately results into a defined cell phenotype. In addition, cellular differentiation represents the path a cell undergoes through multiple stages before reaching a terminally differentiated state and is by nature dynamic. Moreover, epigenetic regulation as well as post-transcriptional control of gene expression are critical for faithful cellular phenotype. Cellular differentiation of progenitor cells into their daughter cells provide a dynamic controllable system to study the epigenetic mechanisms as well as the transcriptional output that take place towards cellular specifications, and the TFs and non-coding RNAs that dictate their differentiation. Here, we have generated time-series transcriptomic and epigenomic data during the differentiation of bone marrow stromal cells towards adipocytes and osteoblasts and characterized a novel approach called EPIC-DREM to construct dynamic GRNs of adipocytes and osteoblasts. In order to focus on shared transcriptional regulators of early commitment of bone marrow stromal cells towards adipocytes and osteoblasts, we have concentrated our analysis on dynamic super-enhancers to prioritize the identified TFs and discovered aryl hydrocarbon receptor (AHR) as a transcriptional regulator of the multipotent state. In addition, the generated of time-series epigenomic data were used as input for linear regression analysis that allowed to predict genes that are dynamically controlled by post-transcriptional regulators such as microRNAs (miRs). Indeed, genes that differ from their predicted expression level as assessed by the residuals of the linear regression model can be informative about their mRNA stability. In order to decipher genes that are under dynamic post-transcriptional control, the standard deviation of gene’s residuals was taken as a dynamic measure of changes in mRNA stability and clustering analysis coupled to microRNA motifs enrichment analysis allowed to identify post-transcriptionally co-regulated mRNAs. Based on the linear regressions analysis, miR-204 was identified as a potential regulator of adipogenesis. Integration of these types of data can contribute to the understanding of transcriptional and post-transcriptional control of cell differentiation and the here established approaches for key regulators identification can be widely applied to study other cell states transitions. [less ▲]

Detailed reference viewed: 25 (4 UL)
Full Text
See detailDeformation Based Curved Shape Representation
Demisse, Girum UL

Doctoral thesis (2017)

Representation and modelling of an objects' shape is critical in object recognition, synthesis, tracking and many other applications in computer vision. As a result, there is a wide range of approaches in ... [more ▼]

Representation and modelling of an objects' shape is critical in object recognition, synthesis, tracking and many other applications in computer vision. As a result, there is a wide range of approaches in formulating representation space and quantifying the notion of similarity between shapes. A similarity metric between shapes is a basic building block in modelling shape categories, optimizing shape valued functionals, and designing a classifier. Consequently, any subsequent shape based computation is fundamentally dependent on the computational efficiency, robustness, and invariance to shape preserving transformations of the defined similarity metric. In this thesis, we propose a novel finite dimensional shape representation framework that leads to a computationally efficient, closed form solution, and noise tolerant similarity distance function. Several important characteristics of the proposed curved shape representation approach are discussed in relation to earlier works. Subsequently, two different solutions are proposed for optimal parameter estimation of curved shapes. Hence, providing two possible solutions for the point correspondence estimation problem between two curved shapes. Later in the thesis, we show that several statistical models can readily be adapted to the proposed shape representation framework for object category modelling. The thesis finalizes by exploring potential applications of the proposed curved shape representation in 3D facial surface and facial expression representation and modelling. [less ▲]

Detailed reference viewed: 64 (36 UL)
Full Text
See detailWohnmobilität in der Großregion – eine interurbane Diskursanalyse mit Fokus auf den Städten Arlon, Thionville und Trier.
Christmann, Nathalie UL

Doctoral thesis (2017)

Effects of residential mobility moulding in uneven development in border regions can be perceived very differently by city councils or planners and the local population. This dissertation focuses on the ... [more ▼]

Effects of residential mobility moulding in uneven development in border regions can be perceived very differently by city councils or planners and the local population. This dissertation focuses on the perceptions of population mobility and dwelling in a transnational cross-border polycentric region in western Europe. The economic development of the Grand Duchy of Luxembourg calls for a constant expansion of the labour market, attracting cross-border commuters and highly mobile professional elites. The concomitant rises in property prices as well as the extreme housing shortages in Luxembourg have led to an expansion of the housing market into the border regions. So far studies have mostly dealt with the socio-demographic characteristics of the transmigrants. This research aims to detect people’s perceptions of the phenomenon by applying a discourse analysis, thus aiming to trigger an increasing awareness for the emerging transnational housing market. [less ▲]

Detailed reference viewed: 13 (7 UL)
See detailArt as an Investment
Nasser Eddine, Ali UL

Doctoral thesis (2017)

During the 1970s and 1980s, the art markets gave abnormal returns. Individuals started speculating on art prices, and institutional investors soon entered the scene. Economists then began evaluating this ... [more ▼]

During the 1970s and 1980s, the art markets gave abnormal returns. Individuals started speculating on art prices, and institutional investors soon entered the scene. Economists then began evaluating this new alternative asset class. In this thesis, we review global art markets, analyze the methodologies employed for studying art as an investment, and seek answers to some fundamental questions. To build solid conclusions, we developed the largest up-to-date dataset of repeat sales of art objects. Our main additional contributions to the literature can be summarized as follows. First, we review and explain the growth in international art markets. Second, we show that it is unreasonable to make a comparison between the two main methodologies used for studying the investment perspective of art: the repeat-sales and hedonic regression frameworks. The returns estimated using the hedonic approach depend greatly on the specifications of the model. Thus, we find that of the two, the repeat-sales models are the most robust. Third, we study the returns on art after accounting for transaction costs. Importantly, we show that taking this fair view renders impractical the widely used art-investment measurement methodologies. Fourth, we revisit the “masterpiece effect”, and find strong evidence supporting its existence. Fifth, we investigate the potential of art investment. We find that the inclusion of art in an optimal portfolio depends significantly on the abnormal returns seen in the 1980s. Omitting these years leads to its exclusion. However, art may add a diversification benefit to an investment portfolio due to its low-to-negative correlation with other asset classes. Sixth, we analyze the optimal holding period of art and find that, in general, the returns increase with the length of the holding period. Nevertheless, we observe significant returns, accompanied with high levels of volatility, for trades made over very short time horizons. We notice that this “flipping” practice has been increasing in recent decades. Finally, we consider the effect some special cases have on art investment returns. We find that artworks that trade frequently tend not to outperform the market. Moreover, the nature of an artwork’s ownership history doesn’t alter returns. We also examine the returns on artworks selected by experts, and find that, surprisingly, they underperform. [less ▲]

Detailed reference viewed: 7 (2 UL)
See detailLa compétence d'incrimination de l'Union européenne. Recherche sur le pouvoir pénal européen
Simon, Perrine UL

Doctoral thesis (2017)

The allocation by the Lisbon Treaty of a genuine criminalisation competence to the Union – article 83 TFEU – prompts the analysis of the existence and the exercice of European criminal law power. It ... [more ▼]

The allocation by the Lisbon Treaty of a genuine criminalisation competence to the Union – article 83 TFEU – prompts the analysis of the existence and the exercice of European criminal law power. It raises the question of the promotion, through criminalisation choices, of essential values to the community. Analysing the criminal law power is interconnected to the question about the nature of the European project as a whole, true existential space of society (ethos) or simple functional space comprised of objectives (telos). Despite the aspiration to clarify the delimitation of European penal power within the new treaty, its ambit remains unclear. An implicit criminalisation competence – an implied criminal law power – could still exist, allowing to overcome the minimum harmonisation provided for by article 83 TFEU. To admit such an implied power would mean revive competence creep. The exercice of the criminalisation competence is progressively framed, beyond the classical principles of subsidiarity and proportionality, by the criminal law principles of ultima ratio as well as the principles of legality and proportionality of criminal offences and penalties expressed in the Charter of fundamental rights. However, these principles have not been taken into account according to the actions of the legislator who appear to follow a securitarian trend. It is to the Court, through an in-depth proportionality check, to determine if the Charter can become the marker of criminal law policy characterised by its moderation and liberalism, and henceforth contribute to a European criminal law identity. [less ▲]

Detailed reference viewed: 12 (3 UL)
Full Text
See detailOn the Use of Alloy in Engineering Domain Specific Modeling Languages
Gammaitoni, Loïc UL

Doctoral thesis (2017)

Domain Specific Modeling Languages (DSMLs) tend to play a central role in modern design processes as they enable the effective involvement of domain experts by focusing on a particular problem domain ... [more ▼]

Domain Specific Modeling Languages (DSMLs) tend to play a central role in modern design processes as they enable the effective involvement of domain experts by focusing on a particular problem domain while abstracting away technical details. In this thesis, we investigate the specification of DSMLs with a particular focus on domain expert driven validation. Mainly, we are interested in developing Alloy-based approaches, allowing the definition of specifications from which instances can be generated and given to the domain experts for the sake of validation. The work we present in this thesis can be divided into three parts: The first part concerns the definition and execution of model transformations defined in Alloy. While Alloy analysis can be used as an execution engine for model transformations, the analysis process is time consuming. Model transformations playing a central role in DSML definitions, the development of a new model transformation language, named F-Alloy, retaining the benefits of Alloy with the added property of being efficiently computable was necessary. The second part focuses on validation. In that domain, our first contribution is a novel approach to the validation of model transformations called Visualization-Based Validation (VBV). VBV relies on the review by domain experts of intuitive depictions of model transformation traces to validate model transformation specifications. The whole process is made efficient by the usage of hybrid analysis, a combination of Alloy analysis and F-Alloy interpretation, allowing to reduce the time needed to analyze model transformations to the time needed to analyze its source. Our second contribution in the validation area is the definition of an Alloy-based approach to the specification and validation of DSMLs and of a design process defining how DSMLs can be validated using Alloy analysis at each iteration of the process. More precisely, we present how the abstract syntax, concrete syntax and operational semantics of a DSML can be defined using Alloy and F-Alloy, and show that the validation of a DSML' s abstract syntax and semantics benefits from the application of its concrete syntax. The third and last part aims at bringing those contributions to the practical world. To achieve this we developed a tool named Lightning implementing the aforementioned contributions. This tool, which belongs to the category of language workbenches, has been successfully used in an inter-disciplinary collaboration to define the Robot Perception System Language (RPSL). Based on this definition of RPSL, a framework has been developed to allow the execution of so called design space explorations. This framework represents a successful application of our approach to the real world problem of having RPSL specifications validated by experts in robotics. [less ▲]

Detailed reference viewed: 34 (6 UL)
Full Text
See detailA Model-Based Framework for Legal Policy Simulation and Compliance Checking
Soltana, Ghanem UL

Doctoral thesis (2017)

Information systems implementing requirements from laws and regulations, such as taxes and social benefits, need to be thoroughly verified to demonstrate their compliance. Several Verification and ... [more ▼]

Information systems implementing requirements from laws and regulations, such as taxes and social benefits, need to be thoroughly verified to demonstrate their compliance. Several Verification and Validation (V&V) techniques, such as reliability testing, and modeling and simulation, can be used for assessing that such systems meet their legal. Typically, one has to model the expected (legal) behavior of the system in a form that can be executed (simulated), subject the resulting models and the system to the same input data, and then compare the observed behavior of the model simulation and system execution. Existing V&V techniques often rely on code and complex logical expressions with no intuitive appeal to legal experts for specifying the expected behavior of a given system. Subsequently, one has no practical way to validate with legal experts that the underlying legal requirements are indeed complete and constitute a faithful representation of what needs to be implemented. Further, manually defining the expected behavior of a system and its test oracles is a tedious and error-prone task. The challenge here is to find a suitable knowledge representation that can be understood by all the involved stakeholders, e.g., software engineers and legal experts, but that remains complete and precise enough to enable automated analysis such as simulation and testing. As real data is seldom accessible in highly regulated domains, V&V requires the generation of synthetic testing data that can be used to build confidence in the reliability of the system under test. In particular, such data has to be structurally and logically well-formed to raise meaningful failures that can help reasoning about the reliability of the system under test. Further, the data should exhibit as much as possible the actual or anticipated system usage to help mimic how the system would behave under realistic circumstances. Generating such data is not a trivial task as the underlying data schemas are usually large and subject to numerous complex domain-related logical constraints. In this thesis, we investigate the use of the Unified Modeling Language (UML) and model-driven technologies, e.g., model to code transformations, to facilitate V&V activities for information systems that have to conform to laws and regulations, while tackling the above challenges. All our technical solutions have been developed and empirically evaluated in close collaboration with a government administration. Concretely, the technical solutions covered by this thesis include: - A modeling notation and methodology for formalizing legal policies. We propose a modeling notation and methodology for building abstract interpretations of the law. Models built using our methodology are simple enough to be understood by the involved stakeholders and are, at the same time, detailed enough to enable automated V&V activities. - A model-based simulation framework. We develop a model-based framework and associated tool support for simulating legal policies, when formalized using the aforementioned modeling methodology. Simulation provides a comparison baseline of how a compliant system should behave. Further, simulation is a mean to support decision-making when considering legal changes. Specifically, we report on a sizable case study where we assess the anticipated economic implications of a given policy change in Luxembourg’s tax law. - A model-based generator of test cases for reliability testing. We develop a heuristic approach for generating valid and representative test cases (data). Our generator is scalable and produces high-quality test data that is suitable for testing the reliability of data-intensive systems, e.g., a tax management system. [less ▲]

Detailed reference viewed: 48 (21 UL)
See detailManaging the City-Region Like a Startup: Entrepreneurialism and Shifting Local Economic Governance in Developing Countries
Fegue, Jean Cyril UL

Doctoral thesis (2017)

The emerging experience of cities in the Global South regarding the complexity of their response to territorial competition's pressures requires the rethinking of the very concept of urban competitiveness ... [more ▼]

The emerging experience of cities in the Global South regarding the complexity of their response to territorial competition's pressures requires the rethinking of the very concept of urban competitiveness. This study proposes the distributionally-sensitive modeling of urban competitiveness (DS-MUC) in a perspective that is driven by the norm of Equity. The DS-MUC is posited as a critical theory to neoliberalism and as a contribution to the social sustainability and to the normative investigation of post-capitalist urban transformations in the Global South.The application of the DS-MUC in the investigation of Da Nang in Vietnam and Cebu City in the Philippines reveals that an interactive, relational and network-based entrepreneurial governance's capacity has a much greater proclivity to deliver Equity and therefore to achieve a 'high-quality competitiveness' than a city's organizing capacity embedded in illiberal, state-paternalistic and public-sector monopolistic arrangements. [less ▲]

Detailed reference viewed: 25 (13 UL)
See detailCorporations and Human Rights: Searching for International Norms for Corporate Conduct in Domestic Case Law
Baglayan Ceyhan, Basak UL

Doctoral thesis (2017)

Recent years have seen much debate concerning the interplay between human rights and corporations. Part of that debate has focused on corporate violations of human rights norms and possible legal ... [more ▼]

Recent years have seen much debate concerning the interplay between human rights and corporations. Part of that debate has focused on corporate violations of human rights norms and possible legal accountability mechanisms for such breaches. The present research is concerned with one such accountability mechanism, namely litigation before domestic courts seeking to enforce corporations’ international obligations and the complaints before the OECD National Contact Points (‘NCPs’). The thesis analyses how domestic courts and the OECD NCPs have conceptualised and implemented corporations’ human rights obligations. It is premised on the assumption that, through their application of international norms in their particular national context, these institutions act to crystallize and clarify the ambit of such norms. [less ▲]

Detailed reference viewed: 19 (4 UL)
Full Text
See detailContextual Integrity and Tie Strength in Online Social Networks: Social Theory, User Study, Ontology, and Validation
Ahmed, Javed UL

Doctoral thesis (2017)

Online Social Networks (OSNs) have become an important part of daily digital interactions for more than half billion users around the world. Unconstrained by physical spaces, OSNs offer to social web ... [more ▼]

Online Social Networks (OSNs) have become an important part of daily digital interactions for more than half billion users around the world. Unconstrained by physical spaces, OSNs offer to social web users new means to communicate, interact, and socialize. Online social networks exhibit many of the characteristics of human societies in terms of forming relationships and sharing personal information. However, current OSNs mainly assume binary, static, and symmetric relationship of equal value between the connected users. In human societies, social relationships are of varying tie strength, dynamic, and asymmetric in nature. The lack of an effective mechanism to represent diversity in social relationships leads to undesirable consequences of users personal information leakage to the unwanted audience and raises privacy concerns. The issue of privacy has received significant attention in both the research literature and the mainstream media. In this dissertation, we conduct a user study to analyze users' attitude towards personal information disclosure in online social networks. The study gives insight into user's information sharing behavior and interaction patterns in online social networks. The findings reveal that personal information disclosure depends on the quality of relationship among the users and it can be easily inferred from user interaction pattern in online social networks. We propose a theoretical framework that addresses the aforementioned issue from a social science perspective and exploits existing social theories of Goffman, Granovetter, and Nissenbaum to model social privacy for OSNs users. Based on this theoretical framework, we developed SOCPRI (SOCial PRIvacy) ontology to represent diversity in social relationships in online social networks. This model regulates personal information disclosure on the basis of the social role and the relationship quality between the OSNs users. The model is evaluated by translating competency questions into description logic (DL) queries to demonstrate the applicability of our approach. The results of ontology evaluation demonstrate the appropriateness of our ontology against proposed requirements. Based on this model a privacy-friendly online social networking environment can be developed to address some of the existing issues such as context collapse and user control. [less ▲]

Detailed reference viewed: 33 (7 UL)
Full Text
See detailModel-Based Specification, Deployment and Adaptation of Robot Perception Systems
Hochgeschwender, Nico UL

Doctoral thesis (2017)

As robots are becoming ubiquitous and more capable, the need for introducing solid robot software development methods is pressing to increase robots' task spectrum. This thesis is concerned with improving ... [more ▼]

As robots are becoming ubiquitous and more capable, the need for introducing solid robot software development methods is pressing to increase robots' task spectrum. This thesis is concerned with improving software engineering of robot perception systems. The presented research employs a model-based approach to provide the means to represent knowledge about robotics software. The thesis is divided into three parts, namely research on the specification, deployment and adaptation of robot perception systems. The first part contributes the design and development of two domain-specific languages, namely RPSL and DepSL. Those languages provide suitable notations and abstractions to enable domain experts to express, compose and explore functional, architectural and deployment design decisions of robot perception systems. The resulting models are interpretable, thus they can be used not only to communicate design decisions to stakeholders, but also to verify them in an early development stage. The second part contributes means for deploying perception systems on real robot systems even in the presence of varying resource conditions. To this end, functional, architectural and deployment models are composed in a graph-structure. Such a graph enables not only humans, but also robots to derive implicitly defined information about their software both at design time and run time. The second part also contributes a reference architecture for deploying robot perception systems. The architecture provides a template solution for integrating not only the models required for deployment, but also all the other means required to carry out deployment. The third part utilizes both RPSL, DepSL and the reference architecture to specify, implement and evaluate three different robot perception systems. Those are capable to satisfy changing requirements induced, for example, by the robot's tasks or environment. This is achieved by proposing algorithms which derive adaptation actions based on models and varying requirements. [less ▲]

Detailed reference viewed: 93 (27 UL)
See detailForm und Funktion des Diminutivs im luxemburgisch-moselfränkischen Übergangsgebiet
Edelhoff, Maike UL

Doctoral thesis (2017)

Diminutive formation is a common word formation process of the Luxembourgish (Lux.) lan-guage and the neighbouring Moselle-Franconian (MsFrc) dialects. The aim of this thesis is to collect and to analyse ... [more ▼]

Diminutive formation is a common word formation process of the Luxembourgish (Lux.) lan-guage and the neighbouring Moselle-Franconian (MsFrc) dialects. The aim of this thesis is to collect and to analyse the characteristics of this process, integrate them into common morpho-logical theories and to establish their geolinguistic properties. On a functional level, the diminutives show the same characteristics in both varieties although the loss of the evaluative meaning is further developed in Lux. than in MsFrc. How-ever, the most apparent differences can be seen on the level of the fomal execution of the di-minutive rule: While the diminutives in MsFrc share many similarities, such as grammatical gender, the singular suffix and to an extent also the plural marking with Standard German, the Lux. language differs greatly from the others. In Lux. the singular formation is quite similar to the one in the other varieties, however, the plural suffix is triggered by the syllabic and prosodic properties of the base noun. Additionally, the grammatical gender of the diminutive is in con-cordance with the gender of the base noun and hence, not influenced by the suffix. The reasons for these peculiar attributes are to be found both in the historical development of the language and its current structural form as well as in its sociolinguistic context. The consequences of the differing structural properties are clearly displayed on a geo-linguistic level. Although the varieties are historically closely related the differing formal ex-pressions of the diminutive meaning lead to the emergence of groups of isoglosses coinciding with the state border. In brief, the history and the present situation of the diminutive serve as evidence that the dialect continuum that once crossed the state border has been falling apart and that it has been replaced by a solid linguistic border that separates the closely related varieties from each other. [less ▲]

Detailed reference viewed: 33 (5 UL)