Hybrid Artificial Intelligence to extract patterns and rules from argumentative and legal texts

Doctoral thesis (2022)

This Thesis is composed of a selection of studies realized between 2019 and 2022, whose aim is to find working methodologies of Artificial Intelligence (AI) and Machine Learning for the detection and ... [more ▼]

This Thesis is composed of a selection of studies realized between 2019 and 2022, whose aim is to find working methodologies of Artificial Intelligence (AI) and Machine Learning for the detection and classification of patterns and rules in argumentative and legal texts. We define our approach as “hybrid”, since different methods have been employed combining symbolic AI (which involves “top-dow” structured knowledge) and sub-symbolic AI (which involves “bottom-up” data-driven knowledge). The first group of these works was dedicated to the classification of argumentative patterns. Following the Waltonian model of argument (according to which arguments are composed by a set of premises and a conclusion), and the theory of Argumentation Schemes, this group of studies was focused on the detection of argumentative evidences of support and opposition. More precisely, the aim of these first works was to show that argumentative patterns of opposition and support could be classified at fine-grained levels and without resorting to highly engineered features. To show this, we firstly employed methodologies based on Tree Kernel classifiers and TFIDF. In these experiments, we explored different combinations of Tree Kernel calculation and different data structures (i.e., different tree structures). Also, some of these combinations employs a hybrid approach where the calculation of similarity among trees is influenced not only by the tree structures but also by a semantic layer (e.g. those using “smoothed” trees and “compositional” trees). After the encouraging results of this first phase, we explored the use of a new methodology which was deeply changing the NLP landscape exactly in that year, fostered and promoted by actors like Google, i.e. Transfer Learning and the use of language models. These newcomer methodologies markedly improved our previous results and provided us with stronger NLP tools. Using Transfer Learning, we were also able to perform a Sequence Labelling task for the recognition of the exact span of argumentative components (i.e. claims and premises), which is crucial to connect the sphere of natural language to the sphere of logic. The last part of this work was finally dedicated to show how to use Transfer Learning for the detection of rules and deontic modalities. In this case, we tried to explore a hybrid approach which combines structured knowledge coming from two LegalXML formats (i.e., Akoma Ntoso and LegalRuleML) with sub-symbolic knowledge coming from pre-trained (and then fine-tuned) neural architectures. [less ▲]

Economics of Migration, Inequalities, and Culture

Doctoral thesis (2022)

The present doctoral thesis consists of three chapters of self-contained works about the economics of migration, inequalities, and culture. In the first chapter, I introduce the thesis outline and discuss ... [more ▼]

The present doctoral thesis consists of three chapters of self-contained works about the economics of migration, inequalities, and culture. In the first chapter, I introduce the thesis outline and discuss each chapter's research questions. The second chapter explores the effects of mass migration on individual attitudes towards migrants. Using several data sources for the mass migration of Ukrainians in Poland between 2014-2016, this chapter is focused on how a massive exogenous increase in the stock of migrant residents and migrant co-workers affects the perception of migrants. Using both an IV methodology and a difference-in-difference analysis, I test two hypotheses: the labor market competition and contact theory and find some evidence favoring the second. First, difference-in-difference analysis shows that Poles become more welcoming to migrants in regions with more job opportunities for migrants. Second, I find that an increase in the size of the migrant group affects attitudes towards migrants positively, inside a group of natives with similar demographic and job skills characteristics. The third chapter explores how poverty can be explained by marital status and gender, using the RLMS-HSE household survey. This research shows that divorced women exhibit lower poverty levels than divorced men by employing longitudinal data from the Russian National Survey (RLMS-HSE) from 2004 to 2019. The result remains qualitatively invariant when considering a theoretical probability to divorce for married couples that take into account the age of the partners, labor force participation, and education. A higher probability to divorce impacts positively only men's poverty level. Investigating an inter-related dynamic model of poverty and labor market participation, we find that divorced women work more than divorced men, which is why divorce hits harder on husbands than on wives. In the fourth chapter of the thesis, we study the effect of past exposure to communist indoctrination during early age (9-14 years) on a set of crucial attitudes in the communist ideology aiming to create the \emph{new communist man/woman}. We focus on the indoctrination received by children during their pioneering years. School pupils automatically became pioneers when they reached 3rd or 4th grade. The purpose of the pioneer years was to educate soviet children to be loyal to the ideals of communism and the Party. We use a regression discontinuity design exploiting the discontinuity in the exposure to pioneering years due to the fall of the USSR in 1991, implying a strong association that hints to causality. We find robust evidence that has been a pioneer has long-lasting effects on interpersonal trust, life satisfaction, fertility, income, and perception of own economic rank. Overall, these results suggest that past pioneers show a higher level of optimism than non-pioneers. Finally, we look for gender differences because various forms of emulation campaigns were used to promote the desired virtues of the new communist woman. However, we find no evidence of the effect of exposure to communism on women. The indoctrination seems to have had more substantial effects on men. [less ▲]

Analysis of urban forms in Europe using radial road network simulations

Doctoral thesis (2022)

Is universal healthcare truly universal? Socioeconomic and migrant inequalities in healthcare

Doctoral thesis (2022)

Through the principle of Universal Healthcare Coverage, many governments across Europe and beyond seek to ensure that all people have equal access to good quality healthcare services, without facing a ... [more ▼]

Through the principle of Universal Healthcare Coverage, many governments across Europe and beyond seek to ensure that all people have equal access to good quality healthcare services, without facing a financial burden. Despite this, studies have highlighted persistent migrant and socio-economic inequalities in the use of healthcare services, and personal health records. Therefore, understanding the complex mechanisms that produce and maintain social inequalities in the effective use of healthcare services is thus an important step towards advancing equity in healthcare. This thesis draws on Bourdieu's forms of capital (cultural, social, economic, and symbolic) to conceptualise and empirically test social inequalities related to healthcare. In doing so, it investigates the factors contributing to socioeconomic and migrant inequalities in the use, navigation and optimisation of healthcare services as well as personal health records. The three studies that make up this thesis empirically test these ideas through statistical modelling on population-based datasets as well as through the analysis of two cross-sectional surveys in Luxembourg and the Greater region. The first study draws on the fifth wave of the Survey of Health, Aging, and Retirement in Europe (SHARE). It used cluster analysis and regression models to explain how the unequal distribution of material and non-material capitals acquired in childhood shape health practices, leading to different levels of healthcare utilisation in later life. The results suggest that although related, both material and non-material capitals independently contribute to health practices associated with the use of healthcare services. The second study used data from a cross-sectional survey to investigate inequalities in the navigation and optimisation of healthcare services, taking into consideration the interplay between perceived racial discrimination and socioeconomic position. It revealed disparities between individuals born in Eastern Europe and the Global South and those born in Luxembourg which were explained by the experience of racial discrimination. It also found that the impact of discrimination on both health service navigation and optimisation was reduced after accounting for social capital. The last study used data from a cross-sectional survey developed as a part of a collaborative project (INTERREG-APPS) to examine the socioeconomic and behavioural determinants in the intention to use personal health record in the Greater region of Luxembourg (Baumann et al., 2020). This study found that people’s desire and actual access to personal health electronic records is determined by different socioeconomic factors, while educational inequalities in the intention to regularly use personal health records were explained by the role of behavioural factors. Taking together, the findings presented in this thesis thus show the value of mobilising Bourdieu’s theoretical framework to understand the mechanisms through which social inequalities in healthcare develop. In addition, it showed the importance of considering racial discrimination when examining migrant, and racial/ethnic differences in health. [less ▲]

Development of gas-generating core-shell amorphous calcium carbonate nanoparticles for contrast-enhanced ultrasound imaging

Doctoral thesis (2022)

Next Generation Mutation Testing: Continuous, Predictive, and ML-enabled

Doctoral thesis (2022)

Software has been an essential part of human life, and it substantially improves production and enriches our life. However, flaws in software can lead to tragedies, e.g. the failure of the Mariner 1 ... [more ▼]

Software has been an essential part of human life, and it substantially improves production and enriches our life. However, flaws in software can lead to tragedies, e.g. the failure of the Mariner 1 Spacecraft in 1962. At the moment, modern software systems are much different from what before. The issue gets even more severe since the complexity of software systems grows larger than before and Artificial Intelligence(AI) models are integrated into software (e.g., Tesla Deaths Report ). Testing such modern artificial software systems is challenging. Due to new requirements, software systems evolve and frequently change, and artificial intelligence(AI) models have non-determination issues. The non-determination of AI models is related to many factors, e.g., optimization algorithms, numerical problems, the labelling threshold, data of the same object but under different collecting conditions or changing the backend libraries. We have witnessed many new testing techniques emerge to guarantee the trustworthiness of modern software systems. Coverage-based Testing is one early technique to test Deep Learning(DL) systems by analyzing the neuron values statistically, e.g., Neuron Coverage(NC) . In recent years, Mutation Testing has drawn much attention. Coverage-based testing metrics can be misleading and easily be fooled by generating tests to satisfy test coverage requirements just by executing the code line. The test suite with one hundred percent coverage may detect no flaw in software. On the contrary, Mutation Testing is a robust approach to approximating the quality of a test suite. Mutation Testing is a technique based on detecting artificial defects from many crafted code perturbations (i.e., mutant) to assess and improve the quality of a test suite. The behaviour of a mutant is likely to be located on the border between correctness and non-correctness since the code perturbation is usually tiny. Through mutation testing, the border behaviour of the subject under test can be explored well, which leads to a high quality of software. It has been generalized to test software systems integrated with DL systems, e.g., image classification systems and autonomous driving systems. However, the application of Mutation Testing encounters some obstacles. One main challenge is that Mutation Testing is resource-intensive. Large resource consumption makes it unskilled in modern software development because the code frequently evolves every day. This dissertation studies how to apply Mutation Testing for modern software systems, exploring and exploiting the usages and innovations of Mutation Testing encountering AI algorithms, i.e., how to employ Mutation Testing for modern software systems under test. AI algorithms can improve Mutation Testing for modern software systems, and at the same time, Mutation Testing can test modern software integrated with DL models well. First, this dissertation adapts Mutation Testing to modern software development, Continuous Integration. Most software development teams currently employ Continuous Integration(CI) as the pipeline where the changes happen frequently. It is problematic to adopt Mutation Testing in Continuous Integration because of its expensive cost. At the same time, traditional Mutation Testing is not a good test metric for code changes as it is designed for the whole software. We adapt Mutation Testing to test these program changes by proposing commit-relevant mutants. This type of mutant affects the changed program behaviours and represents the commit-relevant test requirements. We use the benchmarks from C and Java to validate our proposal. The experiment results indicate that commit-relevant mutants can effectively enhance code change testing. Second, based on the aforementioned work, we introduce MuDelta, an AI approach that identifies commit-relevant mutants, i.e., some mutants that interact with the code change. MuDelta uses manually-designed features that require expert knowledge. MuDelta leverages a combined scheme of static code characteristics as the data feature. Our evaluation results indicate that commit-based mutation testing is suitable and promising for evolving software systems. Third, this dissertation proposes a new approach GraphCode2Vec to learn the general software code representation. Recent works utilize natural language models to embed the code into the vector representation. Code embedding is a keystone in the application of machine learning on several Software Engineering (SE) tasks. Its target is to extract universal features automatically. GraphCode2Vec considers program syntax and semantics simultaneously by combining code analysis and Graph Neural Networks(GNN). We evaluate our approach in the mutation testing task and three other tasks (method name prediction, solution classification, and overfitted patch classification). GraphCode2Vec is better or comparable to the state-of-the-art code embedding models. We also perform an ablation study and probing analysis to give insights into GraphCode2Vec. Finally, this dissertation studies Mutation Testing to select test data for deep learning systems. Since deep learning systems play an essential role in different fields, the safety of DL systems takes centre stage. Such DL systems are much different from traditional software systems, and the existed testing techniques are not supportive of guaranteeing the reliability of the deep learning systems. It is well-known that DL systems usually require extensive data for learning. It is significant to select data for training and testing DL systems. A good dataset can help DL models have a good performance. There are several metrics to guide choosing data to test DL systems. We compare a set of test selection metrics for DL systems. Our results show that uncertainty-based metrics are competent in identifying misclassified data. These metrics also improve classification accuracy faster when retraining DL systems. In summary, this dissertation shows the usage of Mutation Testing in the artificial intelligence era. The first, second and third contributions are on Mutation Testing helping modern software test in CI. The fourth contribution is a study on selecting training and testing data for DL systems. Mutation Testing is an excellent technique for testing modern software systems. At the same time, AI algorithms can alleviate the main challenges of Mutation Testing in practice by reducing the resource cost. [less ▲]

Social Mobility and Inequalities in Health. Bayesian Perspectives on Mortality Risks and Methodological Implications for Statistical Inference.

Doctoral thesis (2022)

Machine Learning-Based Efficient Resource Scheduling for Future Wireless Communication Networks

Doctoral thesis (2022)

The next-generation mobile communication system, e.g., 6G communication system, is envisioned to support unprecedented performance requirements such as exponentially increasing data requests ... [more ▼]

The next-generation mobile communication system, e.g., 6G communication system, is envisioned to support unprecedented performance requirements such as exponentially increasing data requests, heterogeneous service demands, and massive connectivity. When these challenging tasks meet the scarcity of wireless resources, efficient resource management becomes crucial. Conventionally, optimization algorithms, either optimal or suboptimal, are the main approaches for solving resource allocation problems. However, the efficiency of these iterative optimization algorithms can significantly degrade when the problems become large or difficult, e.g., non-convex or combinatorial optimization problems. Over the past few years, machine learning (ML), as an emerging approach in the toolbox, is widely investigated to accelerate the decision-making process. Since applying ML-based approaches to solve complex resource management problems is in its early-stage study, many open issues and challenges need to be solved towards the maturity and practical applications. The motivation and objective of this dissertation lie at investigating and providing answers to the following research questions: 1) How to overcome the shortcomings of extensively adopted end-to-end learning in addressing resource management problems, and which types of features are suited to be learned if supervised learning is applied? 2) What are the limitations and benefits when widely-used deep reinforcement learning (DRL) approaches are used to address constrained and combinatorial optimization problems in wireless networks, and are there tailored solutions to overcome the inherent drawbacks? 3) How to enable ML-based approaches to timely adapt to dynamic and complex wireless environments? 4) How to enlarge the performance gains when the paradigm shifts from centralized learning to distributed learning? The main contributions are organized by the following four research works. Firstly, from a supervised-learning perspective, we address common issues, e.g., unsatisfactory pre- diction performance and resultant infeasible solutions, when end-to-end learning approaches are applied to resource scheduling problems. Based on the analysis of optimal results, we design suited-to-learn features for a class of resource scheduling problems, and develop combined learning-and-optimization approaches to enable time-efficient and energy-efficient resource scheduling in multi-antenna systems. The original optimization problems are mixed-integer programming problems with high-dimensional decision vectors. The optimal solution requires exponential complexity due to the inherent difficulties of the problems. Towards an efficient and competitive solution, we apply fully-connected deep neural network (DNN) and convolutional neural network (CNN) to learn the designed features. The predicted information can effectively reduce the large search space and accelerate the optimization process. Compared to the conventional optimization and pure ML algorithms, the proposed method achieves a good trade-off between quality and complexity. Secondly, we address typical issues when DRL is adopted to deal with combinatorial and non-convex scheduling problems. The original problem is to provide energy-saving solutions via resource scheduling in energy-constrained networks. An optimal algorithm and a golden section search suboptimal approach are developed to serve as offline benchmarks. For online operations, we propose an actor-critic-based deep stochastic online scheduling (AC-DSOS) algorithm. Compared to supervised learning, DRL is suitable for dynamic environments and capable of making decisions based on the current state without an offline training phase. However, for the specific constrained scheduling problem, conventional DRL may not be able to handle two major issues of exponentially-increased action space and infeasible actions. The proposed AC-DSOS is developed to overcome these drawbacks. In simulations, AC-DSOS is able to provide feasible solutions and save around more energy compared to the conventional DRL algorithms. Compared to the offline benchmarks, AC-DSOS reduces the computational time from second-level to millisecond-level. Thirdly, the dissertation pays attention to the performance of the ML-based approaches in highly dynamic and complex environments. Most of the ML models are trained by the collected data or the observed environments. They may not be able to timely respond to the large variations of environments, such as dramatically fluctuating channel states or bursty data demands. In this work, we develop ML-based approaches in a time-varying satellite-terrestrial network and address two practical issues. The first is how to efficiently schedule resources to serve the massive number of connected users, such that more data and users can be delivered/served. The second is how to make the algorithmic solution more resilient in adapting to the time-varying wireless environments. We propose an enhanced meta-critic learning (EMCL) algorithm, combining a DRL model with a meta-learning technique, where the meta-learning can acquire meta-knowledge from different tasks and fast adapt to the new task. The results demonstrate EMCL’s effectiveness and fast-response capabilities in over-loaded systems and in adapting to dynamic environments compare to previous actor-critic and meta-learning methods. Fourthly, the dissertation focuses on reducing the energy consumption for federated learning (FL), in mobile edge computing. The power supply and computation capabilities are typically limited in edge devices, thus energy becomes a critical issue in FL. We propose a joint sparsification and resource optimization scheme (JSRO) to jointly reduce computational and transmission energy. In the first part of JSRO, we introduce sparsity and adopt sparse or binary neural networks (SNN or BNN) as the learning model to complete the local training tasks at the devices. Compared to fully-connected DNN, the computational operations can be significantly reduced, and thus requires less energy consumption and fewer transmitted data to the central node. In the second part, we develop an efficient scheduling scheme to minimize the overall transmission energy by optimizing wireless resources and learning parameters. We develop an enhanced FL algorithm in JSRO, i.e., non-smoothness and constraints - stochastic gradient descent, to handle the non-smoothness and constraints of SNN and BNN, and provide guarantees for convergence. Finally, we conclude the thesis with the main findings and insights on future research directions. [less ▲]

Cognitive Pain Modulation in Young and Older Adults: Understanding the Role of Individual Differences in Frontal Functioning

Doctoral thesis (2022)

Cognitive pain modulation is integral to our quality of life and deeply interwoven with the success of pain treatments but is also characterized by large interindividual variations. Emerging evidence ... [more ▼]

Cognitive pain modulation is integral to our quality of life and deeply interwoven with the success of pain treatments but is also characterized by large interindividual variations. Emerging evidence suggests that one of the driving factors behind these variations are individual differences in frontal functioning. Further evidence indicates that pain-related cognitions, and possibly also emotional distress, may influence the efficacy of pain modulation. Central aim of this project was to assess the role of individual differences in frontal functioning in cognitive pain modulation, with a specific focus on older adults. With respect to this, we also wanted to assess whether individual differences in frontal functions could explain conflicting previous results on age-related changes in the efficacy of cognitive pain modulation. In addition, we wanted to address the role of negative pain-related mindsets and emotional distress on the efficacy of cognitive pain modulation. We tested these research questions across four different studies using two prime paradigms of cognitive pain modulation, namely distraction from pain and placebo analgesia. In Study I, we assessed the role of individual differences in executive functions, emotional distress, and pain-related cognitions in modulating heat pain thresholds in healthy young adults in virtual reality environments with different levels of cognitive load. We found that emotional distress and visuo-spatial short-term memory significantly predicted how participants responded to the low vs high load environment. In Study II, we investigated the role of different forms of cognitive inhibition abilities and negative pain-related cognitions in modulating the efficacy of distraction from (heat) pain by cognitive demand in healthy young adults. We found a significant influence of better cognitive inhibition and selective attention abilities on the size of the distraction effect; however, this association was moderated by the participant’s level of pain catastrophizing, i.e., high pain catastrophizers showed an especially strong association. In Study III, we tested potential age-related differences in distraction from pain in a group of young and older adults while simultaneously acquiring functional brain images. We found no age-related changes at the behavioural level, but a slightly reduced neural distraction effect in older adults. The neural distraction effect size in older adults was furthermore significantly positively related to better cognitive inhibition abilities. In Study IV, we explored potential age-related differences in placebo analgesia in a group of young and older adults (who were partly re-recruited from Study III) while recording their brain activity with an electroencephalogram. Results revealed no age-related differences in the magnitude of the behavioural or electrophysiological placebo response, but older adults showed a neural signature of the placebo effect that was distinct from young participants. Regression analyses revealed that executive functions that showed an age-related decline (as established via group comparisons) were significant predictors of the behavioural placebo response. We furthermore found that better executive functions significantly moderated the association between age group and placebo response magnitude: older adults with better executive functions showed a larger placebo response than young adults whereas worse executive functions were associated with a smaller placebo response, possibly explaining why we found no significant difference at the group level. In summary, all studies provide converging evidence that differences in cognitive functions can significantly affect the efficacy of cognitive pain modulation. Although older adults showed a significant decline in most cognitive functions that we assessed, we found no systematic reduction in the efficacy of cognitive pain modulation (except for a slight reduction in the neural distraction effect size). Closer inspection of the data revealed that older adults may have engaged compensatory mechanisms that enabled them to experience the same (or even higher) level of pain relief as younger adults. We furthermore found evidence for the notion that pain-related cognitions and emotional distress may affect how individuals respond to cognitive pain modulation although this association was less systematic than for cognitive functions. Overall, the present thesis adds to the emerging body of evidence highlighting the importance of executive functions, as indicators of frontal functioning, in cognitive pain modulation. [less ▲]

Understaining and explaining cross-border mobility : a free will / predisposition approach

Doctoral thesis (2022)

This dissertation investigates the drivers of cross-border mobility from a multidisciplinary perspective. Both qualitative and quantitative methodologies are used in order to understand and explain why ... [more ▼]

This dissertation investigates the drivers of cross-border mobility from a multidisciplinary perspective. Both qualitative and quantitative methodologies are used in order to understand and explain why workers cross borders. The major contribution of this dissertation is to highlight new determinants of cross-border mobility such as previous migration experience and health state. These drivers have been disregarded in the literature in the past. Moreover, this dissertation validates the motivations of the workers as a relevant driver of cross-border mobility and provides a state of play of the situation of the cross-border workers in Europe, with a specific focus on French cross- border workers. Firstly, this dissertation provides a review of the explanations of cross- border mobility in the existing literature. Secondly, this dissertation analyses the subjective drivers of cross-border mobility using a qualitative dataset composed of 30 interviews of French workers in Luxembourg collected between January 2018 and May 2019. Results highlight that cross-border workers motivate their decision to commute abroad with financial, professional and personal reasons. Furthermore, the motivations of the cross-border workers vary with respect to their socioeconomic profile. Based on these empirical findings, a model of cross-border labour supply was designed. Thirdly, this dissertation assesses the association between migration capital and cross-border mobility using the French part of the European Labour Force Survey called the Enquête Emploi between 2010 and 2018. Results indicate that migrants commute abroad more than non migrants and are also more likely to do so. Migrant children are more likely to commute abroad, suggesting that the capacity to deal with distance and borders can be transmitted throughout generations. The migration capital is a relevant predictor of commuting behaviour, since the higher the capital endowment, the higher the likelihood is to commute abroad. Additional findings can be mentioned. Internal migration does not increase the likelihood to commute abroad. The acquired migration experience is more useful than the inherited migration experience to be engaged in cross-border mobility. Fourthly, this dissertation examines health disparities between cross-border workers and non cross-border workers using the Enquête Emploi between 2013 and 2018. Results 6 suggest a healthy cross-border phenomenon, the existence of major health disparities among cross-border workers and the rejection of the spillover phenomenon for this specific population. Finally, this dissertation concludes that cross-border mobility is a complex phenomenon still partially explained, probably because of the lack of harmonised dataset about cross-border workers within the EU. Further research on cross- border mobility is needed to better understand this population, especially in public health, where everything remains to be done. [less ▲]

When does finance win? A set-theoretic analysis of the conditions of European financial interest groups' lobbying success on post-crisis bank capital requirements

Doctoral thesis (2022)

Acknowledging the failure of the existing regulatory framework after the global financial crisis of 2008, world leaders vowed to reform financial regulation to strengthen stability and restore trust. The ... [more ▼]

Acknowledging the failure of the existing regulatory framework after the global financial crisis of 2008, world leaders vowed to reform financial regulation to strengthen stability and restore trust. The reform of bank capital requirements was a major item on this agenda: the Group of Twenty (G20) entrusted the reform to the Basel Committee on Bank Supervision (BCBS), whose so-called "Basel framework" constitutes the global standard for the prudential regulation of banking activities. While scholars have highlighted the important concessions that were made to financial interests in this reform, a series of demanding new policy tools—which were strongly opposed by financial industry representatives—were also introduced into the new Basel III framework. This dissertation explores this empirical puzzle and seeks to identify under what conditions European financial interests’ lobbying on the reform of capital requirements was successful, and whether these successes constitute cases of interest group influence. Defining influence as a situation where a proposed reform evolves during the decision-making process (policy shift) in the direction advocated by an actor (lobbying success) and where that evolution is caused by the actors’ lobbying activity vis-à-vis the proposed reform (causal path), this dissertation then considers influence as a multilevel concept, which can be considered present if and only if all three of its components—policy shift, lobbying success and a causal path—are also present. In other words, policy shift, lobbying success and causal paths are the three individually necessary and jointly sufficient conditions for influence, which this study investigates in turn in the case of post-crisis bank capital requirements. The presence or absence of a policy shift is assessed qualitatively by comparing, for twenty-nine policy issues contained in the Basel III framework, the initial BCBS reform proposals with the rules finally enacted at international and European level. The positions of financial and non-financial interest groups on each of these twenty-nine issues are then determined—through a quantitative text analysis of the position papers submitted by interest groups to BCBS and European Commission consultations on Basel III and the CRD and CRR—to determine whether the identified policy shift on a given issue constitutes a case of lobbying success for the interest group. Finally, using fuzzy-set Qualitative Comparative Analysis (fsQCA) to compare in a systematic manner cases in which success is observed and cases where it is absent, I uncover the configurations of conditions sufficient to produce successful lobbying and those sufficient to produce the absence of success, configurations which I then interpret in terms of causal mechanisms. Strong collective action is found, in several forms, to form the basis of causal mechanisms producing successful lobbying. The observed sufficient configurations of conditions however suggest that the causal mechanisms producing success also include key contextual factors that are beyong the control of financial interest groups. The absence of these enabling contextual factors is shown, conversely, to lead to the absence of success. This dissertation contributes to the existing academic literature in several ways. Empirically, first, it adds to the scholarship on bank capital requirements at the international and European level, using novel data to reassess, after the completion of the Basel III reform, the extent to which the final framework meets the initial ambitions. Methodologically, second, this dissertation employs a range of new methods and techniques to take on the challenges of measuring lobbying success and identifying multiple pathways to influence, two fundamental issues for empirical studies of interest group influence. Theoretically, third, the combinatorial approach used here to explore conditions of lobbying success permits an examination of multiple conjunctural causation patterns in interest group influence. [less ▲]

Investigation in reusable composite flooring systems in steel and concrete based on composite behaviour by friction

Doctoral thesis (2022)

The steel-concrete composite systems proved to be very efficient structural solution in terms of material consumption and mechanical response regarding the construction of the structural floor systems ... [more ▼]

The steel-concrete composite systems proved to be very efficient structural solution in terms of material consumption and mechanical response regarding the construction of the structural floor systems whether in the case of industrial and residential buildings or especially in the case of car parks. However, their contemporary application that relies on the utilization of the welded headed studs as a mean to provide the shear connection between the steel section and the concrete chord renders the system unable to be disassembled (in the best case its steel and concrete parts are recycled). Considering the ongoing push from the linear to circular economical models and the application of 3R principle (Reduce, Reuse & Recycle) such systems are unable to furtherly improve their environmental and economic efficiency through reuse schemes. The profound task in this research is development and the verification of the new demountable shear connector solutions that could allow modularity and demountability (hence reusability) of the steel-concrete composite floor systems while retaining their inherent structural advantages. Based on the previous investigations of the demountable shear connector systems (at the first-place bolted solutions) and investigations of mechanical components that were not strictly related to the shear connectors, four demountable shear connector devices were developed. Having in mind the drawbacks of the earlier solutions, adequate detailing and structural measures were applied and the ease of assembly and disassembly was proved on the constructed prototypes. Afterwards, the mechanical properties of devised demountable connector systems were investigated thoroughly through experimental campaign (push tests) and numerical investigation. Based on the experimental and numerical results of the shear connector behaviour it is concluded that the proposed shear connector device Type B possess adequate strength and stiffness and might be considered ductile in accordance with the EN 1994-1-1 allowing for the application of existent design strategies in accordance with the same design code. The force-slip behaviour of the proposed shear connector is explained and adequate analytic model is proposed. Based on the force-slip behaviour model the applicability of the shear connector is verified on a range of composite beams that represent the demountable floor. [less ▲]

Educational and Occupational Differentials in Cognition: Examining the Role of Gender using Population-Based Studies

Doctoral thesis (2022)

Smart cloud collocation: a unified workflow from CAD to enhanced solutions

Doctoral thesis (2022)

Computer Aided Design (CAD) software packages are used in the industry to design mechanical systems. Then, calculations are often performed using simulation software packages to improve the quality of the ... [more ▼]

Computer Aided Design (CAD) software packages are used in the industry to design mechanical systems. Then, calculations are often performed using simulation software packages to improve the quality of the design. To speed up the development costs, companies and research centers have been trying to ease the integration of the computation phase in the design phase. The collocation methods have the potential of easing such integration thanks to their meshless nature. The geometry discretization step which is a key element of all computational method is simplified compared to mesh-based methods such as the finite element method. We propose in this thesis a unified workflow that allows the solution of engineering problems defined by partial differential equations (PDEs) directly from input CAD files. The scheme is based on point collocation methods and proposed techniques to enhance the solution. We introduce the idea of “smart clouds”. Smart clouds refer to point cloud discretizations that are aware of the exact CAD geometry, appropriate to solve a defined problem using a point collocation method and that contain information used to improve locally the solution. We introduce a unified node selection algorithm based on a generalization of the visibility criterion. The proposed algorithm leads to a significant reduction of the error for concave problems and does not have any drawback for convex problems. The point collocation methods rely on many parameters. We select in this thesis parameters for the Generalized Finite Difference (GFD) method and the Discretization-Corrected Particle Strength Exchange (DC PSE) method that we deem appropriate for most problems from the field of linear elasticity. We also show that solution improvement techniques, based on the use of Voronoi diagrams or on a stabilization of the PDE, do not lead to a reduction of the error for all of the considered benchmark problems. These methods shall therefore be used with care. We propose two types of a posteriori error indicators that both succeed in identifying the areas of the domain where the error is the greatest: a ZZ-type and a residual-type error indicator. We couple these indicators to a h-adaptive refinement scheme and show that the approach is effective. Finally, we show the performance of Algebraic Multigrid (AMG) preconditions on the solution of linear systems compared to other preconditioning/solution methods. This family of preconditioners necessitates the selection of a large number of parameters. We assess the impact of some of them on the solution time for a 3D problem from the field of linear elasticity. Despite the performance of AMG preconditions, ILU preconditioners may be preferred thanks to their ease of usage and robustness to lead to a convergence of the solution. [less ▲]

Towards a Unified and Robust Data-Driven Approach. A Digital Transformation of Production Plants in the Age of Industry 4.0

Doctoral thesis (2022)

Nowadays, industrial companies are engaging their global transition toward the fourth industrial revolution (the so-called Industry 4.0). The main objective is to increase the Overall Equipment ... [more ▼]

Nowadays, industrial companies are engaging their global transition toward the fourth industrial revolution (the so-called Industry 4.0). The main objective is to increase the Overall Equipment Effectiveness (OEE), by collecting, storing and analyzing production data. Several challenges have to be tackled to propose a unified data-driven approach to rely on, from the low-layers data collection on the machine production lines using Operational Technologies (OT), to the monitoring and more importantly the analysis of the data using Information Technologies (IT). This is all the more important for companies having decades of existence – as Cebi Luxembourg S.A., our partner in a Research, Development and Innovation project subsidised by the ministry of the Economy in Luxembourg – to upgrade their on-site technologies and move towards new business models. Artificial Intelligence (AI) now knows a real interest from industrial actors and becomes a cornerstone technology for helping humans in decision-making and data-analysis tasks, thanks to the huge amount of (sensors-based) univariate time-series available in the production floor. However, such amount of data is not sufficient for AI to work properly and to make right decisions. This also requires a good data quality. Indeed, good theoretical performance and high accuracy can be obtained when trained and tested in isolation, but AI models may still provide degraded performance in real/industrial conditions. In that context, the problem is twofold: • Industrial production systems are vertically-oriented closed systems that make difficult their communication and their cooperation with each other, and intrinsically the data collection. • Industrial companies used to implement deterministic processes. Introducing AI - that can be classified as stochastic - in the industry requires a full understanding of the potential deviation of the models in order to be aware of their domain of validity. This dissertation proposes a unified strategy for digitizing an industrial system and methods for evaluating the performance and the robustness of AI models that can be used in such data-driven production plants. In the first part of the dissertation, we propose a three-steps strategy to digitize an industrial system, called TRIDENT, that enables industrial actors to implement data collection on production lines, and in fine to monitor in real-time the production plant. Such strategy has been implemented and evaluated on a pilot case-study at Cebi Luxembourg S.A. Three protocols (OPC-UA, MQTT and O-MI/O-DF) are used for investigating their impact on the real-time performance. The results show that, even if these protocols have some disparity in terms of performance, they are suitable for an industrial deployment. This strategy has now been extended and implemented by our partner - Cebi Luxembourg S.A - in its production environment. In the second part of the thesis dissertation, we aim at investigating the robustness of AI models in industrial settings. We then propose a systematic approach to evaluate the robustness under perturbations. Assuming that i) real perturbations - in particular on the data collection - cannot be recorded or generated in real industrial environment (that could lead to production stops) and ii) a model would not be implemented before evaluating its potential deviations, limits or weaknesses, our approach is based on artificial injections of perturbations into the data sets, and is evaluated on state-of-the-art classifiers (both Machine-Learning and Deep-Learning) and data sets (in particular, public sensors-based univariate time series). First, we propose a coarse-grained study, with two artificial perturbations - called swapping effect and dropping effect - in which simple random algorithms are used. This already highlights a great disparity of the models’ robustness under such perturbations that industrial actors need to be aware of. Second, we propose a fine-grained study where instead of testing randomly some parameters' values, we used Genetic Algorithms to look for the models' limits. To do so, we define our multi-objectives optimisation problem with a fitness function as: maximising the impact of the perturbations (i.e. decreasing the most the model's accuracy), while minimising the changes in the time-series (with regards to our two parameters). This can be seen as an adversarial case, where the goal is not to exploit these weaknesses in a malicious way but to be aware of. Based on such a study, methods for making more robust the model and/or for observing such behaviour on the infrastructure could be investigated and implemented if needed. The tool developed in this latter study is therefore ready for being used in a real industrial case, where data sets and perturbations can now be fitted to the scenario. [less ▲]

Data Analysis for Insurance: Recommendation System Based on a Multivariate Hawkes Process

Doctoral thesis (2022)

The objective of the thesis is to build a recommendation system for insurance. By observing the behaviour and the evolution of a customer in the insurance context, customers seem to modify their insurance ... [more ▼]

The objective of the thesis is to build a recommendation system for insurance. By observing the behaviour and the evolution of a customer in the insurance context, customers seem to modify their insurance cover when a significant event happens in their life. In order to take into account the influence of life events (e.g. marriage, birth, change of job) on the insurance covering selection from customers, we model the recommendation system with a Multivariate Hawkes Process (MHP), which includes several specific features aiming to compute relevant recommendations to customers from a Luxembourgish insurance company. Several of these features are intent to propose a personalized background intensity for each customer thanks to a Machine Learning model, to use triggering functions suited for insurance data or to overcome flaws in real-world data by adding a specific penalization term in the objective function. We define a complete framework of Multivariate Hawkes Processes with a Gamma density excitation function (i.e. estimation, simulation, goodness-of-fit) and we demonstrate some mathematical properties (i.e. expectation, variance) about the transient regime of the process. Our recommendation system has been back-tested over a full year. Observations from model parameters and results from this back-test show that taking into account life events by a Multivariate Hawkes Process allows us to improve significantly the accuracy of recommendations. The thesis is presented in five chapters. Chapter 1 explains how the background intensity of the Multivariate Hawkes Process is computed thanks to a Machine Learning algorithm, so that each customer has a personalized recommendation. Chapter 1 is shown an extended version of the method presented in [1], in which the method is used to make the algorithm explainable. Chapter 2 presents a Multivariate Hawkes Processes framework in order to compute the dependency between the propensity to accept a recommendation and the occurrence of life events: definitions, notations, simulation, estimation, properties, etc. Chapter 3 presents several results of the recommendation system: estimated parameters of the model, effects of contributions, backtesting of the model’s accuracy, etc. Chapter 4 presents the implementation of our work into a R package. Chapter 5 concludes on the contributions and perspectives opened by the thesis. [less ▲]

THE INTRICATE LINK BETWEEN DIET, THE HOST IMMUNE SYSTEM, AND THE GUT MICROBIOTA IN THE CONTEXT OF ENTEROPATHOGENIC INFECTIONS AND MULTIPLE SCLEROSIS

Doctoral thesis (2022)

IL-6 Signaling and long non-coding RNAs in liver cancer

Doctoral thesis (2022)

EXCESSIVE MICROBIAL MUCIN FORAGING INDUCED BY DIETARY FIBER DEPRIVATION MODULATES SUSCEPTIBILITY TO INFECTIOUS AND AUTOIMMUNE DISEASES

Doctoral thesis (2022)

The gastrointestinal (GI) mucus layer is a protective and lubricating hydrogel of polymer-forming glycoproteins that covers our intestinal epithelium. This mucus layer serves as an interface between the ... [more ▼]

The gastrointestinal (GI) mucus layer is a protective and lubricating hydrogel of polymer-forming glycoproteins that covers our intestinal epithelium. This mucus layer serves as an interface between the intestinal epithelium and environment as well as a as first line of defense against the potentially harmful microorganisms. While the GI mucus layer closer to the gut epithelium is highly condensed and acts as a physical barrier for invading microorganisms, further away from the epithelium, proteolytic degradation makes it loose. This looser part of the mucus layer serves as an attachment site and a nutrient source for some commensal gut bacteria. The molecular mechanisms that drive the mucus–microbe interactions are emerging and are important to understand the functional role of the gut microbiome in health and disease. Previous work by my research group showed that a dietary fiber-deprived gut microbiota erodes the colonic mucus barrier and enhances susceptibility to a mucosal pathogen Citrobacter rodentium, a mouse model for human Escherichia coli infections. In this PhD thesis, I studied role of the gut mucus layer in the context of various other infectious and autoimmune diseases by inducing the natural erosion of the mucus layer by dietary fiber deprivation. In order to unravel the mechanistic details in the intricate interactions between diet, mucus layer and gut microbiome, I leveraged our previously established gnotobiotic mouse model hosting a synthetic human gut microbiota of fully characterized 14 commensal bacteria (14SM). I employed three different types of infectious diseases for the following reasons: 1) attaching and effacing (A/E) pathogen (C. rodentium), to better understand which commensal bacteria aid in enhancing the pathogen susceptibility when a fiber-deprived gut microbiota erodes the mucus barrier; 2) human intracellular pathogens (Listeria monocytogenes and Salmonella Tyhimurium) to investigate, whether like the A/E pathogen, erosion of the mucus layer could affect the infection dynamics; and 3) a mouse nematode parasite – Trichuris muris, which is a model for the human parasite Trichuris trichiura – to study how changes in the mucin–microbiome interactions drive the worm infection, as mucins play an important role in worm expulsion. In my thesis, I used various combinations of 14SM by dropping out individual or all mucin-degrading bacteria from the microbial community to show that, in the face of reduced dietary fiber, the commensal gut bacterium Akkermansia muciniphila is responsible for enhancing susceptibility to C. rodentium, most likely by eroding the protective gut mucus layer. For my experiments with intracellular pathogens (L. monocytogenes and S. Tyhimurium, I found that dietary fiber deprivation provided protection against the infection by both L. monocytogenes and S. Typhimurium. This protective effect against the pathogens was driven directly by diet and not by the microbial erosion of the mucus layer, since a similar protective effect was observed in both gnotobiotic and germ-free mice. Finally, for the helminth model, I showed that that fiber deprivation-led elevated microbial mucin foraging promotes clearance of the parasitic worm by shifting the host immune response from a susceptible, Th1 type to a resistant, Th2 type. In the context of autoimmune disease, I focused on inflammatory bowel disease (IBD). Although IBD results from genetic predisposition, the contribution of environmental triggers is thought to be crucial. Diet–gut microbiota interactions are considered to be an important environmental trigger, but the precise mechanisms are unknown. As a model for IBD, I employed IL-10-/- mice which are known to spontaneously develop IBD-like colitis in conventional mice. Using our 14SM gnotobiotic mouse model, I showed that in a genetically susceptible host, microbiota-mediated erosion of the mucus layer following dietary fiber deprivation is sufficient to induce lethal colitis. Furthermore, my results show that this effect was clearly dependent on interaction all three factors: microbiome, diet and genetic susceptibility. Leaving out only one of these factors eliminated the lethal phenotype. The novel findings arising from my PhD thesis will help the scientific community to enhance our understanding of the functional role of mucolytic bacteria and the GI mucus layer in shaping our health. Overall, given a reduced consumption of dietary fiber in industrialized countries compared to developing countries, my results have profound implications for potential treatment and prevention strategies by leveraging diet to engineer the gut microbiome, especially in the context of personalized medicine. [less ▲]

ARGUMENT MINING AND ITS APPLICATIONS IN POLITICAL DEBATES

Doctoral thesis (2022)

Presidential debates are significant moments in the history of presidential campaigns. In these debates, candidates are challenged to discuss the main contemporary and historical issues in the country and ... [more ▼]

Presidential debates are significant moments in the history of presidential campaigns. In these debates, candidates are challenged to discuss the main contemporary and historical issues in the country and attempt to persuade the voters to their benefit. These debates offer a legitimate ground for argumentative analysis to investigate political discourse argument structure and strategy. The recent advances in machine learning and Natural Language Processing (NLP) algorithms with the rise of deep learning have revolutionized many natural language applications, and argument analysis from textual resources is no exception. This dissertation targets argument mining from political debates data, a platform rifled with the arguments put forward by politicians to convince a general public in voting for them and discourage them from being appealed by the other candidates. The main contributions of the thesis are: i) Creation, release and reliability assessment of a valuable resource for argumentation research. ii) Implementation of a complete argument mining pipeline applying cutting-edge technologies in NLP research. iii) Launching of a demo tool for argumentative analysis of political debates. The original dataset is composed of the transcripts of 41 presidential election debates in the U.S. from 1960 to 2016. Beside argument extraction from political debates, this research also aims at investigating the practical applications of argument structure extraction, such as fallacious argument classification and argument retrieval. In order to apply supervised machine learning and NLP methods to the data, an excessive annotation study has been conducted on the data and led to the creation of a unique dataset with argument structures composed of argument components (i.e., claim and premise) and argument relations (i.e., support and attack). This dataset includes also another annotation layer with six fallacious argument categories and 14 sub-categories annotated on the debates. The final dataset is annotated with 32,296 argument components (i.e., 16,982 claims and 15,314 premises) and 25,012 relations (i.e., 3,723 attacks and 21,289 supports), and 1628 fallacious arguments. As the methodological approach, a complete argument mining pipeline is designed and implemented, composed of the two main stages of argument component detection and argument relation prediction. Each stage takes advantage of various NLP models outperforming standard baselines in the area, with an average F-score of 0.63 for argument components classification and 0.68 for argument relation classification. Additionally, DISPUTool, an argumentative analysis online tool, is developed as proof-of-concept. DISPUTool incorporates two main functionalities. Firstly, it provides the possibility of exploring the arguments which exist in the dataset. And secondly, it allows for extracting arguments from text segments inserted by the user leveraging the embedded trained model. [less ▲]

Avoiding the Inappropriate: The European Commission and Sanctions under EU Fiscal Policy Coordination

Doctoral thesis (2022)

Since the beginning of the European Economic and Monetary Union, fiscal non-compliance has been subject to the potential imposition of sanctions. However, the extent to which punitive action should be ... [more ▼]

Since the beginning of the European Economic and Monetary Union, fiscal non-compliance has been subject to the potential imposition of sanctions. However, the extent to which punitive action should be automatic – rather than political – is a point of constant discussion among European Union decision-makers. The most recent reform of the Stability and Growth Pact, in the aftermath of the European Sovereign Debt Crisis, has attempted to make sanctions more automatic and has created the possibility to trigger them at earlier stages of the surveillance procedure. With this in mind, the reform has enhanced the powers and autonomy of the European Commission in the application of the new rules. Despite the reinforcement of punitive provisions, the Commission has so far refrained from proposing the imposition of sanctions. Against this background, this thesis aims to answer the question of how we can best explain that the European Commission does not propose financial sanctions because of Member State non-compliance with the Pact’s fiscal objectives. The thesis draws upon four post-crisis cases in which sanctions for fiscal non-compliance might have been imposed – Belgium in 2013, France in 2015, Portugal and Spain in 2016, and Italy in 2018. The thesis uses theory-testing process-tracing methods and applies an adaptation of normative institutionalism that takes into account strategic actor behaviour. Based on this theoretical and methodological framework, it is argued that the normative-strategic minimum enforcement mechanism explains the Commission’s behaviour. Given that the imposition of sanctions is perceived as inappropriate in the cases at hand, Commission actors strategically refrain from applying the enforcement provisions to their full extent. [less ▲]

Optical measurement under space conditions

Doctoral thesis (2022)

The growing interest in space by governmental and private institutions has increased significantly in recent years. The issue of quality control plays an extremely important role in space travel, as ... [more ▼]

The growing interest in space by governmental and private institutions has increased significantly in recent years. The issue of quality control plays an extremely important role in space travel, as possible defects can cause enormous damage. The present work deals with a possible method to improve already existing quality control procedures for space flight. With the help of a 3D scanner, different components are measured and evaluated under space conditions. In particular, the linear thermal expansions are analyzed. The work has shown that the elaborated procedure works for metallic materials. For composites or joints between different materials, positive approaches were shown, which, however, could not be validated within the scope of this work. Components made of pure carbon fiber material cannot be evaluated with the technical equipment used. [less ▲]

Effects of an e-learning platform to improve primary care physicians’ response to domestic violence

Doctoral thesis (2022)

Family Violence (FV) is a broad term that includes different types of violence and abuse that occur within the family, like domestic violence (DV) or Intimate Partner Violence (IPV), Child Abuse or ... [more ▼]

Family Violence (FV) is a broad term that includes different types of violence and abuse that occur within the family, like domestic violence (DV) or Intimate Partner Violence (IPV), Child Abuse or neglect (CA), Elder Abuse (EA) and Female Genital Mutilation (FGM), inter alia. The most prevalent is DV or IPV, declared to be one of the most serious human rights violations, even prior to the aggravated situation brought about by the COVID-19 in 2020. Margaret Chan, 2013 WHO director, called it a global public health care issue of epidemic proportions, but COVID-19 has transformed it into a pandemic as well, earning it the dubious distinction of being the shadow pandemic. Governments’ responses to stop the spread of the infection have forced many families to stay at home, triggering or aggravating cases of IPV and abuse. The prevalence of IPV around the world has been estimated at 30%, although this percentage varies depending on the region of the world or the country, ranging from 20% in the Western Pacific to 33% in the World Health Organisation (WHO) South-East Asia region. Unfortunately, the United Nations Population Fund (UNFPA) predicts that the impact of the pandemic will increase IPV by 20% during the pandemic. Despite the epidemic proportions of this healthcare problem and its enormous and horrendous consequences at all levels (social well-being, physical and mental health), not only for the individual but also their families, IPV remains largely underdiagnosed. Although victims tend to use healthcare services more, and trust health care professionals (HCP) to disclose abuse, they do not do so unless professionals specifically ask for it. Nevertheless, one of the most common barriers that prevent HCP from enquiring is that they do not feel adequately trained to tackle it. Although research on the effectiveness of IPV training for HCP suggests that they improve their knowledge, attitudes, self-perceived readiness to approach this problem, and actual response, this topic has not yet been formally included in the curricula. The World Health Organization and the National Institute for Health have published guidelines for health services and recommendations to facilitate the development and implementation of effective training, underlining the need to improve HCP education. The overall objectives of this thesis are to describe current training provisions on FV in the European Region (FAVICUE), and to investigate the effects of a digital education intervention in improving primary care physicians’ response to DV (E-DOVER). [less ▲]

THERMODYNAMICS OF CHEMICAL ENGINES: A CHEMICAL REACTION NETWORK APPROACH

Doctoral thesis (2022)

Chemical processes in closed systems inevitably relax to equilibrium. Energy can be employed to counteract such tendency and drive reactions against their spontaneous direction. This nonequilibrium ... [more ▼]

Chemical processes in closed systems inevitably relax to equilibrium. Energy can be employed to counteract such tendency and drive reactions against their spontaneous direction. This nonequilibrium driving is implemented in open systems, which living organisms provide the most spectacular examples of. In recent years, experiments in supramolecular chemistry, photochemistry and electrochemistry demonstrated that, by opening synthetic systems to matter and/or energy exchanges with the environment, artificial systems with life-like behaviours can be realized and used to convert energy inputs of different nature into work at both the nanoscopic and the macroscopic level. However, one tool that is still lacking is a firm grasp of the thermodynamics of these chemical engines. In this thesis, we provide it by leveraging the most recent developments of the thermodynamic description of deterministic chemical reaction networks. As main theoretical results, we extend the current theory to encompass nonideal and light-driven systems, thus providing the fundamental tools to treat electrochemical and photochemical systems in addition to the chemically driven ones. We also expand the scope of information thermodynamics to bipartite chemical reaction networks characterized by macroscopic non-normalized concentration distributions evolving in time with nonlinear dynamics. This framework potentially applies to almost every synthetic chemical engine realized until now, and to many models of biological systems too. Here, we undertake the thermodynamic analysis of some of the epitomes in the field of artificial chemical engines: a model of chemically driven self-assembly, an experimental chemically driven molecular motor, and an experimental photochemical bimolecular pump. The thesis provides a thermodynamic level of understanding of chemical engines that is general, complements previous analyses based on kinetics and stochastic thermodynamics, and has practical implications for designing and improving synthetic systems, regardless of the particular type of powering or chemical structure. [less ▲]

Older people's use of assistive technologies and alternative means to cope with health-related declines

Doctoral thesis (2022)

Health-related declines are risk factors for autonomous living and quality of life of older people. Yet, they have various action possibilities (e.g., use of technical and personal support, health ... [more ▼]

Health-related declines are risk factors for autonomous living and quality of life of older people. Yet, they have various action possibilities (e.g., use of technical and personal support, health services, environmental adaptation) which may reduce these risks. Sadly, the use rates of these options are far from perfect, so that their positive effects are not sufficiently realized. With this in mind, the main goal of this dissertation was to gain a better understanding of the factors that influence older people's use of the various action possibilities to cope with illness, functional decline, activity limitations, and participation restrictions. Five studies using different but complementary methods pursued this goal. Study I provided a systematic review of 12 empirical studies on the effectiveness and use of self-care assistive technologies (ATs). It found that self-care ATs are efficient with respect to reduced care hours and increased independence level. The use of such technologies was associated with three different kinds of factors including personal, contextual, technological device factors. Study II conducted a systematic review of 23 theoretical models of AT use from an action-theoretical perspective on lifespan development. It revealed that these models considered a limited range of internal and external context factors of AT use. None of them contained perceived discrepancies between the actual and desired developmental situation, goals to reduce these discrepancies, action alternatives to AT use, and decision-making about AT use by other persons or a joint decision-making. Study III contained a qualitative meta-synthesis of 7 qualitative studies on subjective reasons for the use or nonuse of ATs. It thus considered a branch of research that is independent from the ones covered by Study 1 and 2 and it used a different method. It found 25 subjective reasons referring to user’s beliefs and desires of which 18 were not contained in AT use models. However, they could be included in more comprehensive models to increase their predictive value. Study IV focused on the construction of an “Actional Model of older people´s coping with health related declines” for explaining the use of 8 major action possibilities (7 beyond ATs). Its development followed a recent theory construction methodology and recent principles for constructing a practically useful theory. It integrated results from the studies 1, 2 and 3 as well as from other relevant literature. Central explanatory variables are perceived discrepancies between actual and desired development, discrepancy reduction and prevention goals and internal as well as external context factors. Outcome variables are the 8 courses of action and their results. Study V examined the view of experts (professional caregivers of older people) regarding central components of the Actional Model of Older people´s Coping with Health-Related Declines developed in study 4. Theory generating expert interviews were conducted to further clarify key components of the model. The results led to their further specification, such as the contents of discrepancy reduction and prevention goals, further motivating and demotivating goals, and external context factors as barriers and facilitators of their use. The findings of the dissertation are discussed with respect to the advancement of empirical, theoretical and methodological knowledge. Implications for future research and for the improvement of practical applications in gerontological case management and developmental counselling are highlighted. [less ▲]

A posteriori error estimation for finite element approximations of fractional Laplacian problems and applications to poro–elasticity

Doctoral thesis (2022)

This manuscript is concerned with a posteriori error estimation for the finite element discretization of standard and fractional partial differential equations as well as an application of fractional ... [more ▼]

This manuscript is concerned with a posteriori error estimation for the finite element discretization of standard and fractional partial differential equations as well as an application of fractional calculus to the modeling of the human meniscus by poro-elasticity equations. In the introduction, we give an overview of the literature of a posteriori error estimation in finite element methods and of adaptive refine- ment methods. We emphasize the state–of–the–art of the Bank–Weiser a posteriori error estimation method and of the adaptive refinement methods convergence results. Then, we move to fractional partial differential equations. We give some of the most common discretization methods of fractional Laplacian operator based equations. We review some results of a priori error estimation for the finite element discretization of these equations and give the state–of–the–art of a posteriori error estimation. Finally, we review the literature on the use of the Caputo’s fractional derivative in applications, focusing on anomalous diffusion and poro-elasticity applications. The rest of the manuscript is organized as follow. Chapter 1 is concerned with a proof of the reliability of the Bank–Weiser estimator for three–dimensional problems, extending a result from the literature. In Chapter 2 we present a numerical study of the Bank–Weiser estimator, provide a novel implementation of the estimator in the FEniCS finite element software and apply it to a variety of elliptic equations as well as goal-oriented error estimation. In Chapter 3 we derive a novel a posteriori estimator for the L2 error induced by the finite element discretization of fractional Laplacian operator based equations. In Chapter 4 we present new theoretical results on the convergence of a rational approximation method with consequences on the approximation of fractional norms as well as a priori error estimation results for the finite element discretization of fractional equations. Finally, in Chapter 5 we provide an application of fractional calculus to the study of the human meniscus via poro-elasticity equations. [less ▲]

Verifiable, Secure and Privacy-Preserving Computation

Doctoral thesis (2022)

In this thesis, I present the research I conducted with my co-authors on numerous areas of verifiable, secure, and privacy-preserving computation during my doctoral studies at the University of Luxembourg ... [more ▼]

In this thesis, I present the research I conducted with my co-authors on numerous areas of verifiable, secure, and privacy-preserving computation during my doctoral studies at the University of Luxembourg, where Professor Peter Ryan advised me. In the first part, I study the functional encryption scheme. In the standard setting of functional encryption, it is assumed both the Central Authority (CA) and the encryptors to run their respective algorithms faithfully. However, in the case of dishonest parties, the security of the cryptosystem may be violated. It means that dishonest parties can cause inconsistent results which may not be detected. In the first part, we improve on this situation by considering Inner-Product Encryption (IPE), a special case of functional encryption and a primitive that has attracted wide interest from practitioners and researchers in the last decade. Specifically, we construct the first efficient verifiable Inner Product Encryption (VIPE) scheme according to the inner-product functionality. As the next step, we construct a verifiable IPE that satisfies unconditional verifiability, whereas privacy relies on the standard assumption. The second part of this thesis presents my research on e-voting protocols. I revisit the coercion-resistant e-voting protocol by Juels, Catalano and Jakobsson (JCJ) and, particularly, the attempts to make it usable and practical. In JCJ the user needs to handle cryptographic credentials and fake these in case of coercion. We present a hardware-independent protocol that can be implemented using a combination of a digitally stored cryptographic length key and a PIN only known by the voter. The long credential could be stored in several places or hidden via steganography. At the ballot casting phase, the software will input the digital key and the password to form the credential submitted with the vote. Depending on the level of coercion, the coerced voter can either fake the long credential or, for stronger levels of coercion, the voter can reveal the digitally stored credential to the coercer but fake the PIN. Due to our improved tally, the coercer will not know if he got faked credentials or PINs. On the other hand, since the voter memories the PIN is a high chance of users making a PIN typo error which will invalidate the vote and remain undetected. Note that naively giving feedback on the correctness of the PIN is not possible for coercion-resistance as it would allow the coercer to check whether he got a fake PIN or not. Instead, we will define a set of allowed PIN errors (e.g., chosen by the election administrator). We will consider a ballot valid if it has a correct PIN or an allowed PIN error but invalid for other PINs. At the tally phase, we construct protocols that secretly check whether a given PIN is in the set of allowed PINs and will sort out invalid ballots. We also design another End-to-End verifiable e-voting scheme achieving coercion-resistance via deniable vote updating. We propose a new e-voting system that enables voters with an intuitive mechanism to update their possibly coerced vote in a deniable way. What is more, our e-voting system does not introduce any additional trust assumptions for end-to-end verifiability and vote privacy besides the standards. Moreover, we demonstrate that our e-voting system can be instantiated efficiently for practical elections. With these properties, our e-voting system has the potential to close the gap between theory and practice in coercion-resistant e-voting. [less ▲]

State Responsibility for Judicial Acts in Investment Arbitration

Doctoral thesis (2022)

This thesis deals with the issue of state responsibility for judicial acts in investment arbitration. In general, it is undisputed that states can be held internationally liable for the conduct of their ... [more ▼]

This thesis deals with the issue of state responsibility for judicial acts in investment arbitration. In general, it is undisputed that states can be held internationally liable for the conduct of their judiciary. The protection against denial of justice was traditionally the main international standard used for determining state responsibility for judicial acts, characterized by a relatively high threshold. Only with the proliferation of primary obligations of states towards individuals, the practical relevance of this standard was somewhat decreased. Nonetheless, this type of state responsibility – arising from domestic judicial acts – remains particularly sensitive due to the special nature and position of the national judiciary. It is commonly noted that role of national courts and their decision-making process is so particular that international courts and tribunals should not treat national judicial acts in an identical manner as acts of other state organs. Against this background, this thesis examines whether and to what extent the investment tribunals’ practice on cases concerning domestic judicial acts differs from comparable practices of other international adjudicatory bodies – in particular, the International Court of Justice and the European Court of Human Rights. Through the analysis of various standards of protection – when applied to judicial acts – and the relevant case law, the thesis aims to address two different questions. Firstly, whether the standards of protection granted investment treaties substantially differ from standards commonly applied by other international adjudicatory bodies. Secondly, if the investment tribunals practice’ in cases concerning host states’ judicial acts is influenced only by these substantive standards or protection, or it is also shaped by more general features of investment law and the investor investor-state dispute settlement mechanism. Ultimately, this research – through the analysis of these particular types of investor-state disputes – tries to offer novel insights into the deficiencies of the ISDS regime that should be taken into account, especially in light of the current initiatives for the reform of this regime. [less ▲]

Population structure and phenotypical traits of Campylobacter jejuni circulating in Luxembourg

Doctoral thesis (2022)

Campylobacter is the leading cause of bacterial gastroenteritis worldwide. The most prevalent species, C. jejuni, is a strict microaerobic, capnophilic, and thermotolerant pathogen. Given its growth ... [more ▼]

Campylobacter is the leading cause of bacterial gastroenteritis worldwide. The most prevalent species, C. jejuni, is a strict microaerobic, capnophilic, and thermotolerant pathogen. Given its growth requirements, the ability of C. jejuni to persist in food environments and be transmitted throughout food processing has long puzzled scientists. This study aimed to compare the different genetic profiles of C. jejuni strains, isolated in Luxembourg, at the core genome (cg) and whole genome (wg) levels to elucidate its genetic population structure. Using phenotypical assays in controlled conditions and functional genomics analyses from wgMLST data, the study also investigated the possible link between phenotypic traits and emergence or persistence of genotypes. A high concordance in strain clustering was observed between genomic lineage classifications and the epidemic and endemic signals, regardless the three cgMLST typing schemes used. The higher genome stability within genomic lineages supports the hypothesis of a clonal expansion with monomorphic patterns over time and sources. A high correlation was observed between phenotypes and host-specific or generalist clonal complexes for oxidative stress, adhesion to inert surfaces, biofilm formation, and acclimation to aerobic conditions responses. These data allowed the establishment of metaphenotypes specific to the genomic lineages. Functional genomic analysis revealed factors that may contribute to the spatiotemporal survival of recurrent strains. These results also suggest the selection of better-adapted and persistent C. jejuni strains to environmental stresses throughout the transmission to human. [less ▲]

Automation of Controller Area Network Reverse Engineering: Approaches, Opportunities and Security Threats

Doctoral thesis (2022)

Controller Area Network (CAN ) is the de-facto in-vehicle communication system in the automotive industry today. CAN data represents a valuable source of information regarding the vehicle, which can be ... [more ▼]

Controller Area Network (CAN ) is the de-facto in-vehicle communication system in the automotive industry today. CAN data represents a valuable source of information regarding the vehicle, which can be exploited for a multitude of purposes by aftermarket companies, from fleet management to infotainment. With the rise of Vehicular Ad Hoc Networks (VANETs) and autonomous driving, we can expect the amount of data transiting on the CAN bus to further augment in the near future. While not encrypted, the communication inside the CAN bus is typically encoded using proprietary formats of the Original Equipment Manufacturers (OEM s) in order to prevent easy access to the information exchanged on the network. However, given the unwillingness of the OEM s to disclose the formats of most of the CAN signals of commercial vehicles (cars in particular) to the general public, the most common way to obtain such information is through reverse engineering. Recently, researchers have started investigating the automation of this process to make it faster, scalable and standardised. Aside from the evident advantages that it would bring to the industry, the automation of CAN bus reverse engineering has also gained interest in the scientific community, where automotive cybersecurity is a prominent topic. While achieving convincing results, the automation of CAN reverse engineering is still invasive, often includes complex hardware configurations or requires the presence of a human operator in the vehicle. This dissertation aims to analyse the main advancements achieved in the field of CAN bus reverse engineering and shed light on open issues. In the first part of this dissertation, we explore opportunities and challenges of the automation of CAN bus reverse engineering and present three approaches that achieve different degrees of automation. The first, FastCAN, is based on the taxonomy of signals. Its goal is to provide a complete, standardised and modular pipeline for semi-automated reverse engineering and reduce the total time for data collection. The second, CSI, is a Machine Learning (ML )-based algorithm for the identification of critical signals working under limited assumptions. We use CSI as a case study to investigate whether CAN reverse engineering can be achieved with no other hardware than a dongle for the collection of raw data. The third, CANMatch, is a complete and fully automated approach based on frame matching. Through CANMatch we seek to demonstrate that the reuse of CAN frame IDs can be exploited to reverse engineer a high number of signals with minimal hardware requirements and human effort. In the second part of this dissertation, we discuss the implications that the full automation of the reverse engineering process has on the security of the bus. In this context, we investigate whether the anonymisation of the CAN frame IDs is sufficient to prevent frame-matching based reverse engineering. The results highlight that ML models can fingerprint CAN frames despite the anonymisation of their IDs. Finally, we propose a defence against frame fingerprinting based on traffic mutations, such as padding on the payload and morphing on the sending frequency. We conclude that traffic mutations are a promising study direction to prevent frame-matching based reverse engineering. [less ▲]

Machine learning force fields: towards modelling flexible molecules

Doctoral thesis (2022)

Accurate modelling of chemical and physical interactions is crucial for obtaining thermodynamic and dynamical properties of any chemical system, enabling a myriad of possible applications. Many of these ... [more ▼]

Accurate modelling of chemical and physical interactions is crucial for obtaining thermodynamic and dynamical properties of any chemical system, enabling a myriad of possible applications. Many of these applications are computationally prohibitive when using advanced Computational Chemistry (CompChem) methods even on modern supercomputers. Because of this, machine learning (ML) force fields (FFs), combining the accuracy of state-of-the-art ab initio methods and the efficiency of classical FFs, are being increasingly used to reconstruct potential-energy surfaces (PESs) of molecules and solids. It is precisely the synergy of ML and CompChem that has revolutionized the field in the last decade, rising the applications to a qualitatively new level. Despite this great success, there are still many unsolved challenges. In this context, my thesis aims to investigate the capability of the existing MLFFs to provide simultaneously accurate and efficient models offering unprecedented insights into the (thermo)dynamics of realistic molecular systems. Using the examples of molecular interactions that are pervasive in (bio)chemical systems, we show a counterintuitive effect of strengthening of such interactions, as well as an unexpected prevalence of quantum nuclear fluctuations over thermal contributions at room temperature. We reveal that, when dealing with complex PESs, the predictions of state-of-the-art ML models (BPNN, SchNet, GAP, and sGDML) greatly depend on the descriptor used, and on the region of the PES being sampled. Given the varying performance of MLFFs, we present a descriptor optimization scheme improving simultaneously the accuracy and efficiency of ML models. Our results show that the commonly employed strategies followed to construct both local and global descriptors need to be improved because the optimal descriptors are a non-trivial combination of local and global features. Therefore, the work presented in this thesis highlights the potential of MLFFs to provide insights into chemical systems while, at the same time, discloses the current limitations preventing the construction of accurate MLFFs for more realistic systems. Also, I propose the optimization of the description of interactions within an ML model as a valuable step towards obtaining efficient and accurate MLFFs of large and flexible molecules. Overall, this thesis suggests that the full workflow for building ML models still needs significant elaboration. Despite this finding, the combination of CompChem and ML methods in atomistic modelling promises to enable us to solve multiple problems in different areas, such as medicine, materials design, pharmacology, energy production, environmental sciences, among others. [less ▲]

Detection of Sentiment in Luxembourgish User Comments

Doctoral thesis (2022)

Sentiment is all around us in everyday life. It can be found in blog posts, social media comments, text messages and many other places where people express themselves. Sentiment analysis is the task of ... [more ▼]

Sentiment is all around us in everyday life. It can be found in blog posts, social media comments, text messages and many other places where people express themselves. Sentiment analysis is the task of automatically detecting those sentiments, attitudes or opinions in written text. In this research, the first sentiment analysis solution for the low-resource language, Luxembourgish, is conducted using a large corpus of user comments published on the RTL Luxembourg website www.rtl.lu. Various resources were created for this purpose to set the foundation for further sentiment research in Luxembourgish. A Luxembourgish sentiment lexicon and an annotation tool were built as external resources that can be used for collecting and enlarging training data for sentiment analysis tasks. Additionally, a corpus of mainly sentences of user comments was annotated with negative, neutral and positive labels. This corpus was furthermore automatically translated to English and German. Afterwards, diverse text representations such as word2vec, tf-idf and one-hot encoding were used on the three versions of the corpus of labeled sentences for training different machine learning models. Furthermore, one part of the experimental setup leveraged linguistic features for the classification process in order to study their impact on sentiment expressions. By following such a broad strategy, this thesis not only sets the basis for sentiment analysis with Luxembourgish texts but also intends to give recommendations for conducting sentiment detection research for other low-resource languages. It is demonstrated that creating new resources for a low-resource language is an intensive task and should be carefully planned in order to outperform working with translations where the target language is a high-resource language such as English and German. [less ▲]

Dissecting Complex Microglia Heterogeneity in Neurodegeneration

Doctoral thesis (2022)

Architectural Support for Hypervisor-Level Intrusion Tolerance in MPSoCs

Doctoral thesis (2022)

Increasingly, more aspects of our lives rely on the correctness and safety of computing systems, namely in the embedded and cyber-physical (CPS) domains, which directly affect the physical world. While ... [more ▼]

Increasingly, more aspects of our lives rely on the correctness and safety of computing systems, namely in the embedded and cyber-physical (CPS) domains, which directly affect the physical world. While systems have been pushed to their limits of functionality and efficiency, security threats and generic hardware quality have challenged their safety. Leveraging the enormous modular power, diversity and flexibility of these systems, often deployed in multi-processor systems-on-chip (MPSoC), requires careful orchestration of complex and heterogeneous resources, a task left to low-level software, e.g., hypervisors. In current architectures, this software forms a single point of failure (SPoF) and a worthwhile target for attacks: once compromised, adversaries can gain access to all information and full control over the platform and the environment it controls, for instance by means of privilege escalation and resource allocation. Currently, solutions to protect low-level software often rely on a simpler, underlying trusted layer which is often a SPoF itself and/or exhibits downgraded performance. Architectural hybridization allows for the introduction of trusted-trustworthy components, which combined with fault and intrusion tolerance (FIT) techniques leveraging replication, are capable of safely handling critical operations, thus eliminating SPoFs. Performing quorum-based consensus on all critical operations, in particular privilege management, ensures no compromised low-level software can single handedly manipulate privilege escalation or resource allocation to negatively affect other system resources by propagating faults or further extend an adversary’s control. However, the performance impact of traditional Byzantine fault tolerant state-machine replication (BFT-SMR) protocols is prohibitive in the context of MPSoCs due to the high costs of cryptographic operations and the quantity of messages exchanged. Furthermore, fault isolation, one of the key prerequisites in FIT, presents a complicated challenge to tackle, given the whole system resides within one chip in such platforms. There is so far no solution completely and efficiently addressing the SPoF issue in critical low-level management software. It is our aim, then, to devise such a solution that, additionally, reaps benefit of the tight-coupled nature of such manycore systems. In this thesis we present two architectures, using trusted-trustworthy mechanisms and consensus protocols, capable of protecting all software layers, specifically at low level, by performing critical operations only when a majority of correct replicas agree to their execution: iBFT and Midir. Moreover, we discuss ways in which these can be used at application level on the example of replicated applications sharing critical data structures. It then becomes possible to confine software-level faults and some hardware faults to the individual tiles of an MPSoC, converting tiles into fault containment domains, thus, enabling fault isolation and, consequently, making way to high-performance FIT at the lowest level. [less ▲]

ESSAYS ON ECONOMIC INEQUALITY IN RUSSIA

Doctoral thesis (2022)

There can be little doubt regarding the importance of inequality problem in Russia. Within the scope of this thesis, I have focused on documenting and understanding trends in inequality in Russia from ... [more ▼]

There can be little doubt regarding the importance of inequality problem in Russia. Within the scope of this thesis, I have focused on documenting and understanding trends in inequality in Russia from 1994-2015. The four chapters analyze different aspects of inequality in Russia. The second chapter of the thesis aims at documenting and explaining changes in income inequality in Russia from 1994-2015. First, I provide evidence for inequality and poverty trends in Russia. Second, I provide a detailed examination of the main determinants of the above-mentioned changes. I document that changes in socio-demographic characteristics together with labour market employment of individuals have not affected changes in income inequality, poverty and income levels. Changes in income sources, in particular earnings and pensions, are the key drivers of changes. A fall in income inequality and poverty and increase in income levels is the result of increasing real levels in pensions and public sector earnings and a decreasing dispersion of private sector earnings. The third chapter focuses on the two dimensions of inequality: income and consumption. First, I provide evidence on the inequality trends in income and consumption. Second, I uncover evidence on the joint analysis in income and consumption inequality. The analysis reveals large differences in consumption across income levels and significant differences in consumption changes given income. I document that low and high income top savers have not changed their consumption behavior. The fourth chapter of this thesis addresses the issue of adjusting income to regional price differences by suggesting an approach to estimating cost-of-living indices without data on prices. The proposed approach relies on household survey data only. The key idea of this approach is that the cost-of-living indices between regions are reflected in average differences in income among homogenous individuals. Homogenous individuals experience similar financial needs, which are captured by observed individual characteristics, and the same level of satisfaction with economic conditions. A matching technique is applied to pair individuals from different regions. Finally, the last chapter focuses on the relationship between income changes and political trust in Russia by answering the following question: Do income changes matter for political trust? The paper shows that income changes do not matter for political trust. Furthermore, neither income gains nor income losses play a significant role in the relationship between income changes and trust. I, however, find that short-run income changes from 2012-2011 are positively associated with political trust and that this link is stronger for lower incomes. [less ▲]

Oncometabolite 2-hydroxygluterate regulates gene expression through preferential inhibition of specific histone demethylases

Doctoral thesis (2022)

In several types of cancer but also in some cases of 2-hydroxyglutaric aciduria, a gain of function mutation in the IDH enzymes leads to the accumulation of 2-hydroxyglutarate. The oncometabolite 2 ... [more ▼]

In several types of cancer but also in some cases of 2-hydroxyglutaric aciduria, a gain of function mutation in the IDH enzymes leads to the accumulation of 2-hydroxyglutarate. The oncometabolite 2-hydroxyglutarate has been identified as inhibitor of TET and JmjC-domain containing histone demethylases due to its structural similarities with a-KG, the required co-factor for these demethylases. Accumulation of 2-hydroxyglutarate can lead to changes in DNA and histone methylation contributing to the initiation and progression of cancer. The effects of 2-hydroxyglutarate on repressive histone markers such as histone H3 lysine 27 and histone H3 lysine 9 methylation have been highlighted in several studies although the effects on markers of open chromatin such histone H3 lysine 36 and histone H3 lysine 4 methylation remain unclear. Additionally, in studies investigating the effects of 2-hydroxyglutarate, it has been difficult to distinguish if the pathological effects are attributed to DNA or histone methylation changes. Saccharomyces cerevisiae could serve as a model organism to study the effects of 2-hydroxyglutarate specifically on histone markers as this model is devoid of DNA methylation. In this study we show that elevated 2-hydroxyglutarate levels in S. cerevisiae can lead to genetic background-dependent gene expression changes that are accompanied by altered H3K4 and H3K36 methylation only at specific and often unrelated loci. By working with histone demethylase knockouts strains we show that while inhibition of all of the H3K4 and H3K36 demethylase contributes to the observed methylation changes, only the preferential inhibition of Rph1 is sufficient to induce extensive gene expression changes. These results provide novel insights into genome-wide effects of 2-hydroxyglutarate and highlight Rph1, the yeast homolog of KDM4 demethylases, as its preferential demethylase target. [less ▲]

Vun iwwerfëlltene Bussen bis bei déi beschte Witzer. Morphologische Variation im Luxemburischen – Eine variations- und perzeptionslinguistische Studie.

Doctoral thesis (2022)

Die vorliegende Arbeit beschäftigt sich mit morphologischer Variation im Luxemburgischen und verfolgt dabei zwei Hauptziele: Erstens sollen ausgewählte, zu einem großen Teil noch unerforschte Phänomene ... [more ▼]

Die vorliegende Arbeit beschäftigt sich mit morphologischer Variation im Luxemburgischen und verfolgt dabei zwei Hauptziele: Erstens sollen ausgewählte, zu einem großen Teil noch unerforschte Phänomene auf Grundlage eines umfangreichen Sprachdaten-Korpus systematisch beschrieben und die Variation, die diese Phänomene prägt, detailliert analysiert werden. Zweitens will die Forschungsarbeit mithilfe eines maßgeschneiderten Testverfahrens anschließend eruieren, wie diese Variationsphänomene von Luxemburgisch-Sprecher.innen wahrgenommen und bewertet werden. Die Dissertation liefert somit nicht nur einen wichtigen Beitrag dazu, morphologische Variation und somit auch Wandel im Luxemburgischen besser zu verstehen, sondern auch die Etablierung der perzeptionslinguistischen Forschungsdimension innerhalb der Luxemburgistik voranzutreiben. [less ▲]

Quantifying some properties of curves and arcs on hyperbolic surfaces

Doctoral thesis (2022)

Motivated by the ergodicity of geodesic flow on the unit tangent bundle of a closed hyperbolic surface and its applications, this thesis includes three parts: Part 1. We present a type of quantitative ... [more ▼]

Motivated by the ergodicity of geodesic flow on the unit tangent bundle of a closed hyperbolic surface and its applications, this thesis includes three parts: Part 1. We present a type of quantitative density of closed geodesics and orthogeodesics on complete finite-area hyperbolic surfaces. The main results are upper bounds on the length of the shortest closed geodesic and the shortest doubly truncated orthogeodesic that are Y-dense on a given compact set on the surface. Part 2. We investigate the terms arising in Luo-Tan’s identity, namely showing that they vary monotonically in terms of lengths and that they verify certain convexity properties. Using these properties, we deduce two results. As a first application, we show how to deduce a theorem of Thurston which states, in particular for closed hyperbolic surfaces, that if a simple length spectrum "dominates" another, then in fact the two surfaces are isometric. As a second application, we show how to find upper bounds on the number of pairs of pants of bounded length that only depend on the boundary length and the topology of the surface. This is joint work with Hugo Parlier and Ser Peow Tan. Part 3. Inspired by a number theoretic application of Bridgeman’s identity, the combinatorial proof of McShane’s identity by Bowditch and its generalized version by Labourie and Tan, we describe a tree structure on the set of orthogeodesics and give a combinatorial proof of Basmajian’s identity in the case of surfaces. We also introduce the notion of orthoshapes with associated identity relations and indicate connections to length equivalent orthogeodesics and a type of Cayley-Menger determinant. As another application, dilogarithm identities following from Bridgeman’s identity are computed recursively and their terms are indexed by the Farey sequence. [less ▲]

DAXX maintains heterochromatin at (peri)centromeres independently of ATRX, preventing R-loop-induced DNA double strand breaks and genomic instability

Doctoral thesis (2022)

A TALE OF ADVERSITY: THE IMPACT OF EARLY LIFE INFECTION BEYOND THE IMMUNE SYSTEM

Doctoral thesis (2022)

Early life adversity (ELA) is associated with a higher risk for chronic and noncommunicable diseases in adulthood. Changes in the HPA axis and the immune system have been proposed to underlie this ... [more ▼]

Early life adversity (ELA) is associated with a higher risk for chronic and noncommunicable diseases in adulthood. Changes in the HPA axis and the immune system have been proposed to underlie this association, however current ELA research remains mainly focused on neglect, abuse and low socioeconomic status as sources of childhood adversity. Early-life adversity covers a range of physical, social and environmental stressors. Acute viral infections in early life are a major source of such adversity and have been associated with a broad spectrum of later-life effects outside the immune system or “off-target”. These include an altered hypothalamus-pituitary-adrenal axis and metabolic reactions. Here, we used a murine postnatal day 14 (PND 14) Influenza A (H1N1) infection model and applied a semi-holistic approach including phenotypic measurements, gene expression arrays and diffusion neuroimaging techniques to investigate HPA axis dysregulation, energy metabolism, brain connectivity, microbial composition and immune cell shift . We then extended our mouse model for a postnatal day 56 (PND56) viral challenge to study immune reactivation after an early life “priming” event. We could show that an early life influenza A virus infection led to an immediate shift in lung microbiota and long-term changes in gut microbiome composition. We also observed several sex-specific effects: retarded growth of males, baseline blood glucose levels being increased in females and decreased in males, and baseline corticosterone levels were reduced while total number of CD3+ cells were higher in males. At the same time, we found a microbial composition shift persistent several weeks after the early life infection. For the brain, MRI scans identified reduced connectivity in the cortex, midbrain and cerebellum which were accompanied by tissue-specific gene expression signatures. Early-life infection appeared to have independently affected each of the systems, potentially independent of HPA axis or immune perturbations. [less ▲]

Refugees’ Integration Process in Luxembourg: The case of Arab refugees’ post-political transformations in the Arab region

Doctoral thesis (2022)

2011 was significant in the Arab countries; it started a people's political movement in many countries. It was known as (The Arab Spring). This Arab spring led to instability and insecurities in many ... [more ▼]

2011 was significant in the Arab countries; it started a people's political movement in many countries. It was known as (The Arab Spring). This Arab spring led to instability and insecurities in many countries, which resulted in a large flow of asylum seekers to neighbouring countries and Europe. This flow reached its peak in 2015, and many of them ended up arriving in Luxembourg. The Arab-speaking population is relatively tiny in Luxembourg compared to other neighbouring countries. Still, numbers increased from 1200 residents in 2011 to over 7000 due to this flow of asylum seekers, mainly from Syria and Iraq. While Luxembourg had a specific demographic structure and a multilingual context compared to other European countries, the refugees faced a different situation concerning their integration into Luxembourgish society. Multilingualism is one of the main challenges asylum seekers face, especially if they do not have previous experience or competencies in any European language. Language learning played a prominent role in allowing people to find job opportunities, decent housing, and be independent of the state and social support. In this dissertation, I am trying to investigate how the integration process is functioning for this specific group of people in Luxembourg, which factors play a role in their integration, and how the support measures are valid. I used a qualitative research approach with data collected through semi-structured interviews with asylum seekers who had already received their refugee status to analyse their perception and understanding of their integration process. The interviews were conducted in their native language (Arabic), which gave me access to a more straightforward free discussion with the interviewees but added a limitation of the need to translate the selected excerpts and not having the possibility to translate the whole interviews. The main results are that the Arab refugees’ integration process had several aspects. Concerning the prejudgment before arrival, the waiting time of their asylum application, then after holding the refugee status, other elements that they faced played a role as a barrier; or a challenge for their integration processes like housing, language learning, job market and family conditions. I concluded that different components are needed to fulfil their integration needs, and they can be structured in organized stages for integration. The multiculturalism of Luxembourg can play a positive role in facilitating integration but, simultaneously, can create challenges for people to adapt and progress in their integration process. The diverse components can be organized in a tower model built on the different integration components to reach the level needed for good integration. [less ▲]

MOBILE APPLICATION FOR PHYSICAL ACTIVITY: DEVELOPMENT OF A THEORY‐INFORMED, PERSONALIZED MHEALTH INTERVENTION FOR ADOLESCENTS (MAPA)

Doctoral thesis (2022)

The beneficial impact of physical activity (PA) has been extensively documented, with well-evidenced improved physical and mental health across lifespan, together with increased life expectancy. Yet, a ... [more ▼]

The beneficial impact of physical activity (PA) has been extensively documented, with well-evidenced improved physical and mental health across lifespan, together with increased life expectancy. Yet, a lack of PA and increased sedentary lifestyle continue to represent a serious public health burden. Insufficient levels of physical inactivity have also been observed in adolescents, which is alarming, as PA levels may be transferred into adulthood. Given the ubiquitous use of smartphones by adolescents, these devices may offer feasible means to reach young populations and deliver an intervention aimed at increasing PA participation. The aim of the current project, therefore, concerns the development, delivery and pilot-evaluation of a smartphone application intervention for adolescents 16-18 who are insufficiently active, in order to promote PA engagement and decrease sedentary time. Overall, three independent studies were conducted. Within the first study, a scoping literature review was performed. The results present a range of evidence (both quantitative and qualitative) available on smartphone-based mHealth PA interventions as well as the development and evaluation trajectory of mHealth PA interventions and systematic theory- and evidence-based practices and methods that are implemented along this trajectory. Within the second study, a focus group discussion was conducted and features and components that are preferred by adolescents (aged 16-18) in apps promoting PA were identified. Finally, within our third study, the development and evaluation of the MAPA intervention were reported. The MAPA within-subject trial demonstrated that smartphone-based intervention produced significant reductions in sedentary time in adolescents during the first week of the trial. This trend, while remaining positive, diminished over time. Our findings indicate that there was no effect of the intervention on MET-based MVPA minutes, although the descriptive increase may give reason for further investigation. Although the results suggested no overall change in heart rate based MVPA minutes, the results from the change point analyses suggest that the personalized PA prompts significantly increased HR per minute (bpm) during the second week of the study. There were no significant increases in participants’ overall step count; however, the personalized PA prompts resulted in a marginally significant increased step counts per minute in the second week of the study. These results suggest the feasibility and promise of smartphone-based PA interventions with personalized PA suggestions for adolescents. This study also provides first information and an example for researchers and practitioners on how to guide development of future smartphone-based mHealth PA interventions in adolescents. Future investigations should focus on the replication of these findings and testing the potential of scalability of such an intervention within larger population samples. [less ▲]

On Trustworthy AI and Localized Complex Network Analytics

Doctoral thesis (2022)

We live in a world where the interaction of many different entities results in the formation of complex systems. The communication between billions of smart devices, interactions of millions of people in ... [more ▼]

We live in a world where the interaction of many different entities results in the formation of complex systems. The communication between billions of smart devices, interactions of millions of people in social networks, and the existence of our biological life, which is based on seamless interactions between hundreds of genes and proteins within our cells, all are just a few examples of the complex systems surrounding us. At the core of these complex systems, there is clear evidence of a complex network, which symbolizes the interaction between the system’s components. Analytical metrics and algorithms derived from graph theory are used in network analysis to understand the functionality of complex systems, anticipate system behavior, and control changes. Many of these global patterns in complex networks are generally influenced by decisions made by communities. Communities are a tightly connected group of nodes with sparse connections to the rest of the network. These modular structures are crucial to understanding the complex network due to being closely tied to the system’s functional and topological features. They can, for example, represent modules of proteins with similar functionality in a protein interaction network or influence dynamic network activities such as opinion and epidemic propagation. Local community detection methods have gained popularity among other strategies to dis- cover communities in a complex network. The traditional methods are based on a top-down approach acquiring global information about the entire network; however, due to the growing size and complexity of existing networks, they often result in tangled communities, hence not providing functional information of the network. The primary goal of this thesis is to provide methods and solutions for local network analysis. The following components comprise the thesis contribution: 1) Introduce a transformation approach to construct networks from relational data and describe how network structure affects community detection, 2) Provide a comprehensive evaluation of current local community detection techniques and suggest a locality exploration scheme (LES) for community detection algorithms, 3) Develop a local community detection Algorithm (LCDA) and employ it on real-world data, 4) Extend LCDA to LCDA-GO, which integrates biological functional information and detects protein communities in the cell on the PPI network. Thereby, this thesis proposes a novel community detection algorithm that addresses the shortcomings of prior algorithms by presenting a local method. The applicability of the suggested algorithm is investigated by running it on real-world PPI networks. 7 Furthermore, this thesis contributes to industrial technical reports and whitepapers on the standardization and regulation of big data and Artificial Intelligence (AI). The thesis addresses the critical issues of digital trust in big data and AI by incorporating technical standardization and cutting-edge research solutions. The key contributions include, but are not limited to: 1) A comprehensive review of the state-of-the-art in numerous scientific and standard materials regarding privacy and trustworthiness concerns, including the introduction of privacy leaks and mitigation measures in big data and AI, 2) Investigating the societal implications of artificial intelligence standards in light of the recently initiated worldwide and European standardized processes, 3) Design and implementation of a scheme that connects scientific contributions and stan- dardization efforts in the direction of AI conformity assessment. The contribution of the thesis to standards demonstrates an impact on both scientific and standardization communities by contributing to both and offering recent outcomes from each. [less ▲]

Decision-Making Processes and Certainties in Heritage Reconstructions Using the Example of Larochette Castle, Luxembourg

Doctoral thesis (2022)

To reconstruct a historical building is to make countless decisions, weighing often conflicting sources, dealing with gaps in the data and adapting to new information. This thesis investigates how ... [more ▼]

To reconstruct a historical building is to make countless decisions, weighing often conflicting sources, dealing with gaps in the data and adapting to new information. This thesis investigates how individual aspects of a reconstruction can be classified by their degree of accuracy or the certainty that they are correct, how the public understands - and can be taught to understand - these distinctions, and how this system can be applied to existing reconstructions. [less ▲]

Design, Cryptanalysis and Protection of Symmetric Encryption Algorithms

Doctoral thesis (2022)

This thesis covers results from several areas related to symmetric cryptography, secure and efficient implementation and is divided into four main parts: In Part II, Benchmarking of AEAD, two articles will ... [more ▼]

This thesis covers results from several areas related to symmetric cryptography, secure and efficient implementation and is divided into four main parts: In Part II, Benchmarking of AEAD, two articles will be presented, showing the results of the FELICS framework for Authenticated encryption algorithms, and multiarchitecture benchmarking of permutations used as construction block of AEAD algorithms. The Sparkle family of Hash and AEAD algorithms will be shown in Part III. Sparkle is currently a finalist of the NIST call for standardization of lightweight hash and AEAD algorithms. In Part IV, Cryptanalysis of ARX ciphers, it is discussed two cryptanalysis techniques based on differential trails, applied to ARX ciphers. The first technique, called Meet-in-the-Filter uses an offline trail record, combined with a fixed trail and a reverse differential search to propose long differential trails that are useful for key recovery. The second technique is an extension of ARX analyzing tools, that can automate the generation of truncated trails from existing non-truncated ones, and compute the exact probability of those truncated trails. In Part V, Masked AES for Microcontrollers, is shown a new method to efficiently compute a side-channel protected AES, based on the masking scheme described by Rivain and Prouff. This method introduces table and execution-order optimizations, as well as practical security proofs. [less ▲]

Large graphene oxide flakes: self-assembly and electro-optical switching in liquid crystal phase

Doctoral thesis (2022)

Graphene oxide (GO) flakes in aqueous suspension self-organize into liquid crystal (LC) phases. They have been studied for various applications because of their unique properties, but mostly in their bi ... [more ▼]

Graphene oxide (GO) flakes in aqueous suspension self-organize into liquid crystal (LC) phases. They have been studied for various applications because of their unique properties, but mostly in their bi-phasic suspensions since high concentrations was needed for their pure liquid crystal. The stability of a suspension of GO flakes over time is essential for their application. A crucial change in GO biphasic suspension is the isotropic - liquid crystal phase separation. Here I study the isotropic-nematic phase transition for large GO flakes with an average size of ~38 µm for 570 days to examine their stability. I found that large GO flakes make pure LC phases at a very low concentration of 0.7 mg/mL (0.035 vol%), completely stable over time. I observed that the equilibrium concentration for making a LC phase increases over time, and it is not constant until it reaches a critical concentration for the pure nematic phase. As typical LCs, GO LCs exhibits birefringence property and, interestingly, low-field-induced birefringence. Thus, graphene-based lyotropic LCs are an attractive class of materials for electro-optic devices. To control the electro-optic properties of these high-performance functional materials at a macroscopic scale, it is necessary to understand their self-assembly individually. I study this here by determining the large GO flakes arrangement and spatial order in the nematic phase via synchrotron small-angle X-ray scattering and direct visualization of the flakes with fluorescence confocal laser scanning microscopy. The high fluorescence of large GO flakes makes it possible to individually study the assembly of flakes in real-time when suspended in aqueous dispersions. I successfully performed an enhanced electro-optical switching of large GO flakes in their low concentration nematic phase by using a low electric field, unlike what has been reported so far. I study the optical behavior of GO LCs using polarizing optical microscopy in static and dynamic conditions to follow the flake reassembly under the application of the electric field. [less ▲]

Kummer theory for commutative algebraic groups

Doctoral thesis (2022)

This thesis consists of four research articles that treat different aspects of Kummer theory for commutative algebraic groups, with particular emphasis on explicit and effective results.

From Smart to Secure Contracts: Automated Security Assessment and Improvement of Ethereum Smart Contracts

Doctoral thesis (2022)

Modern blockchains, such as Ethereum, gained tremendously in popularity over the past few years. What partially enables this large increase are so-called smart contracts. These are programs that are ... [more ▼]

Modern blockchains, such as Ethereum, gained tremendously in popularity over the past few years. What partially enables this large increase are so-called smart contracts. These are programs that are deployed and executed across the blockchain. However, just like traditional programs, smart contracts are subject to programming mistakes. Although, unlike traditional programs their code is publicly available and immutable. Hence, as smart contracts become more popular and carry more value, they become a more interesting target for attackers. In the past few years, several smart contracts have been exploited, resulting in assets worth millions of dollars being stolen. In this dissertation, we explore the security of smart contracts from three different perspectives: vulnerabilities, attacks, and defenses, and demonstrate that, as so often, "smart" does not imply "secure". In the first part of the dissertation, we study the automated detection of vulnerabilities in smart contracts, without requiring prior access to source code. We start by building a symbolic execution framework for detecting integer bugs that leverages taint analysis to reduce false positives. However, as symbolic execution is notorious to produce false positives, we explore hybrid fuzzing as an alternative. As a result, we propose a hybrid fuzzer for smart contracts that combines symbolic execution with fuzz testing and leverages data dependencies across state variables to efficiently generate transaction sequences. Our approach is capable of detecting more vulnerabilities with less false positives. In the second part of the dissertation, we explore various ways to mount attacks against smart contracts. We start by proposing a framework to detect and quantify classical smart contract attacks (e.g., reentrancy, integer overflows, etc.) on past transactions by combining logic-driven and graph-driven analysis. Afterwards, we study the effectiveness of a new type of fraud known as honeypots, by scanning the entire blockchain for different types of honeypots using symbolic execution. Next, we present a methodology to measure the prevalence of so-called frontrunning attacks, which follow from the rise of decentralized finance and the sharp increase of users trading on decentralized exchanges. Our results show that attackers are making a fortune by manipulating the order of transactions. In the third and final part of the dissertation, we discuss several defense mechanisms for smart contracts. We first propose a solution that developers can use to automatically patch vulnerable smart contract bytecode using context-sensitive patches that dynamically adapt to the bytecode that is being patched. However, this does not solve the issue of already deployed smart contracts. To that end, we present a second solution that enables security experts to write attack patterns that are triggered whenever malicious control and data flows are detected. Once a transaction is detected to be malicious, all state changes are rolled back and the attack is thereby prevented. These attack patterns are written using a domain-specific language and are managed via a smart contract. The latter enables decentralization, guarantees the distribution of security updates, and warrants transparency. [less ▲]

Understanding the role of diet and microbiome in colorectal cancer

Doctoral thesis (2022)

FABRICATION AND INVESTIGATION OF PHYSICOCHEMICAL AND STRUCTURAL GRADIENTS FOR BIOSENSING

Doctoral thesis (2022)

Controlling biomolecule-surface interactions with nano- and micro-engineered surfaces is of great interest in biomedical applications such as tissue regeneration and biosensing platforms. Developing high ... [more ▼]

Controlling biomolecule-surface interactions with nano- and micro-engineered surfaces is of great interest in biomedical applications such as tissue regeneration and biosensing platforms. Developing high-performance functional bio-interfaces for cell-surface or protein-surface interactions necessitates optimizing the interface by modifying material surface variables. Surface gradients are a category of combinatorial technique that enables monitoring and high-throughput optimization of biomolecule-surface interactions by providing a gradually varying surface parameter(s) on a small scale and across an extended area length. It is elaborated that a surface gradient not only greatly reduces the required time and labour of conducting numerous separate experiments for producing several distinct samples but also minimises the inter-batch errors associated with. In this context, multigradients are particularly promising for advanced bio-interface optimisation since they incorporate two or more separate gradients that evolve independently across different directions. While gradients have been vastly studied in past two decades, reporting different surface gradients of chemistry, topography, or mechanical nature in either nano or larger scales, there have been few studies on multigradients, due to the limited operational flexibility required for generating more than one gradient on the surface. First, plasma technologies were explored for establishing a suitable fabricating method for generating spatial variation of surface chemistry along a direction. Both the mask-assisted static and maskless dynamic deposition were examined via two different plasma technologies, namely atmospheric pressure plasma and low-pressure plasma. Depending on the electrical conductivity of the chosen substrates and the nature of the coatings, different surface characterisations were performed on the generated samples. Surface chemistry, surface morphology and wettability properties of the treated surfaces were mainly investigated. As a result, two chemistry gradients were reported; first, an oxygen-functional chemistry gradient deposited with a single-step approach via a programmed corona discharge based on the polymerisation of HMDSO with varying flow rates of oxygen. The chemistry gradient consisted of 7 deposition conditions spanning between mostly organic and inorganic coating also exhibiting the surface energy gradient along a polyethylene foil with length of 10 cm. The surface morphology was also altered as oxygen level was increasing, leading to mild gradual surface roughening. Second, a nitrogen-functional chemistry gradient with the specific feature of enhanced water stability was reported via polymerisation of ethylene with gradually varying ammonia flow rates using a mask-assisted static deposition approach with low pressure capacitively coupled radio frequency plasmas. A smooth coating exhibiting a chemistry gradient consisted of four deposition conditions, and a subsequent surface energy gradient was achieved along 1 cm width of a 2x1cm Si chip. Following that, a versatile experimental setup was presented for developing the next class of surface gradients, the structural or topography gradients, which benefited from a rational design and soft lithography. As a result, a total of 4 topography gradients were reported, two of which were stochastic density gradients and the other two being periodical nanocluster density and periodical size gradients. The gradient was formed based on time-dependent incubation of the functionalised material surface with the chosen precursor and electrostatic interactions between the two. The main experimental inputs were the precursor flow rate, dimension of the experiment chamber and dimension of the substrate. For material surface functionalisation, various classes of chemistries were employed, including aminosilane monolayers, cross-linked plasma polymer, and copolymer templates for developing either stochastic or periodic arrangements of the surface features. The kinetics of incubation of each functional surface was monitored with real-time QCM before gradient formation allowing a prediction of surface coverage and all the generated gradients were investigated for their surface morphology. The obtained micrographs and the respective experimental plots and theoretical fittings confirmed the successful formation of stochastic and periodical topography gradients. Surface-enhanced Raman spectroscopy (SERS) studies revealed the high potential of gold nanocluster density gradients for SERS-based biosensing applications. However, despite exceptionally strong SERS signals recorded on the nanoparticle density gradient (generated on the plasma polymer template), the SERS response diminished at some spots along the surface, revealing a noncontinuous SERS variation. Meanwhile, gold domes did not demonstrate any enhancement as a function of size variation. Wettability analyses were performed selectively on the stochastic gold nanoparticle density gradient utilizing both the experimental sessile drop method and theoretical modelling to investigate the probable wetting regime. The theoretical modelling indicated good agreement with the experimental WCAs and indicated Wenzel, full wetting regime.As the ultimate objective, an orthogonal surface gradient was presented. The approach was based on depositing the previously reported nitrogen-functional chemistry gradient in a perpendicular direction over the unidirectional stochastic gold nanoparticle density gradient. As confirmed by XPS and ToF-SIMS, the surface chemical composition was retained after coating and did not change due to the presence of the underlying conductive gold nanoparticle layer. The surface morphology was significantly altered after being coated with the top plasma layer, demonstrating an overall decreased roughness variation compared to the unidirectional nanoparticle density gradient. Furthermore, the surface wettability variation was significantly lower when compared to the wettability variation scale of the integrated unidirectional gradients. [less ▲]