Was Luxemburgs Grundschulen richtig machen: Ein Value-Added-Vergleich im luxemburgischen Schulkontext

Poster (2023, September 18)

Kurz-Abstract (120 Wörter) Luxemburgs Bildungssystem ist geprägt von multi-kulturellen und vielsprachigen Schüler:innen und einem zweimaligen Wechsel der Instruktionssprache. Dies führt zu sehr ... [more ▼]

Kurz-Abstract (120 Wörter) Luxemburgs Bildungssystem ist geprägt von multi-kulturellen und vielsprachigen Schüler:innen und einem zweimaligen Wechsel der Instruktionssprache. Dies führt zu sehr unterschiedlichen Voraussetzungen für die Schullaufbahn der Schüler:innen. Das Ziel des vorliegenden SIVA-Projekts (Systematic Identification of High Value-Added in Educational Contexts) ist herauszufinden, welche pädagogischen Strategien Schulen mit hohen Value-Added (VA)-Werten für Schuleffektivität anwenden und was andere Schulen von ihnen lernen können, um diese Ungleichheiten abzubauen. Zuerst ermittelten wir 16 Schulen, die stabil hohe, mittlere oder niedrige VA-Werte aufwiesen. Danach sammelten wir Daten anhand von Fragebögen und Unterrichtsbeobachtungen über pädagogische Strategien und das Schulklima und glichen sie mit repräsentativen Schulmonitoringergebnissen ab. Wir werden das SIVA-Projekt, seine Ziele und die Datenerhebung diskutieren, die zu unserem reichhaltigen Datensatz aus sechs Perspektiven führte. Zusammenfassung (480 Wörter) In einem multi-kulturellen und vielsprachigen Land wie Luxemburg können leicht Bildungsungleichheiten entstehen. Unterschiedliche zu Hause gesprochene Sprachen, Migrationshintergründe oder der sozioökonomische Status einer Familie können zu ungleichen Erfolgschancen in der Schule werden. Gepaart mit einem Schulsystem, in dem zweimal die Instruktionssprache gewechselt wird, führt diese Vielfalt zu unterschiedlichen Voraussetzungen für das Erlernen von Mathematik und Sprachen und prägt somit die Schullaufbahn der Schüler:innen (Hadjar & Backes, 2021). Diese Gemengelage ist einerseits herausfordernd für Schüler:innen, Lehrkräfte und Schulen, zeigt aber andererseits, dass es gelingende soziale und pädagogische Praktiken geben muss, diese Herausforderungen zu meistern, da die Schulen weiterhin effektiv arbeiten. In den USA wurde Schuleffektivität häufig mit Value-Added-Werten (VA) quantifiziert, welche durch ihre Instabilität zu ungerechtfertigten Finanzierungs- und Personalentscheidungen führten (Emslander, Levy, Scherer, et al., 2022). Ziel des Projekts Systematic Identification of High Value-Added in Educational Contexts (SIVA; Emslander, Levy, & Fischbach, 2022) ist es, dieses repressiv genutzte Instrument der VA-Werte konstruktiv anzuwenden. VA ist ein statistisches Regressionsverfahren, um die Effektivität von Schulen unter Berücksichtigung unterschiedlicher Schüler:innenhintergründe gerecht zu schätzen. Wir untersuchten, (1) was hocheffektive Schulen "richtig" machen und (2) was andere Schulen von ihnen lernen können, um Ungleichheiten abzubauen. In Zusammenarbeit mit der Section Qualité Scolaire des Observatoire National de l’Enfance, de la Jeunesse et de la Qualité Scolaire, untersuchten wir die Unterschiede zwischen Schulen mit stabil hohen, mittleren oder niedrigen VA-Werten aus verschiedenen Perspektiven. Zunächst haben wir 16 Schulen ermittelt, die über zwei Jahre hinweg stabile hohe, mittlere oder niedrige VA-Werte aufwiesen. Als Zweites sammelten wir Fragebogen- und Unterrichtsbeobachtungsdaten über ihre pädagogischen Strategien, den Hintergrund der Schüler:innen und das Schulklima. Als Drittes glichen wir unsere Daten mit den Ergebnissen des luxemburgischen Schulmonitorings ÉpStan (LUCET, 2021) ab. Wir haben die Variablen auf der Grundlage von Lernmodellen ausgewählt, die sich auf Aspekte wie die Schulorganisation oder das Klassenmanagement konzentrieren (z.B. Hattie, 2008; Klieme et al., 2001). Darüber hinaus untersuchten wir die Besonderheiten des luxemburgischen Schulsystems, die in internationalen schulischen Lernmodellen nicht vertreten sind (z. B. die Einteilung in zweijährige Lernzyklen, die mehrsprachige Schulumgebung und die vielfältige Schülerschaft). Wir werden das SIVA-Projekt, seine Ziele und Besonderheiten diskutieren, die zu Daten aus 49 Klassenzimmerbeobachtungen und Fragebögen mit über 500 Zweitklässler:innen, ihren Eltern, 200 Lehrkräften sowie Schulleiter:innen und Schulaufsichtsbehörden führte. Literature Emslander, V., Levy, J., & Fischbach, A. (2022). Systematic Identification of High “Value-Added” in Educational Contexts (SIVA). https://doi.org/10.17605/OSF.IO/X3C48 Emslander, V., Levy, J., Scherer, R., & Fischbach, A. (2022). Value-added scores show limited stability over time in primary school. PLOS ONE, 17(12), e0279255. https://doi.org/10.1371/journal.pone.0279255 Hadjar, A., & Backes, S. (2021). Bildungsungleichheiten am Übergang in die Sekundarschule in Luxemburg. https://doi.org/10.48746/BB2021LU-DE-21A Hattie, J. (2008). Visible Learning: A synthesis of over 800 meta-analyses relating to achievement (0 ed.). Routledge. https://doi.org/10.4324/9780203887332 Klieme, E., Schümer, G., & Knoll, S. (2001). Mathematikunterricht in der Sekundarstufe I: “Aufgabenkultur” und Unterrichtsgestaltung. TIMSS - Impulse für Schule und Unterricht, 43–57. LUCET. (2021). Épreuves Standardisées (ÉpStan). https://epstan.lu [less ▲]

Migration and Health in Older Adults: The Cognitive and Health Profile of Lusophone Immigrants in Luxembourg in Comparison to Local Samples from Portugal

Poster (2023, September 04)

Objective: This study aimed to investigate differences in cognition, mental health, and general health among older Lusophone immigrants (>55 years old) in Luxembourg compared to a sample residing in ... [more ▼]

Objective: This study aimed to investigate differences in cognition, mental health, and general health among older Lusophone immigrants (>55 years old) in Luxembourg compared to a sample residing in Portugal. Methods: Data from immigrants (N=83, mean age=62.08, SD=6.511, 72.84% women) were sourced from the MEDITAGING study, funded by the Luxembourg National Research Fund (FNR)-Project Reference 15240063. Portuguese data were collected by the GABT (N=55, mean age=68.18, SD = 5.869, 74.55% women) and PREDICTOR projects (N=83, mean age=67.86, SD=5.24, 83.13% women). The outcomes were the Body Mass Index and the score on the Mini-Mental State Examination, the Trail Making Test (TMT), the 5-Item Geriatric Depression Scale (GDS-5), and the Geriatric Anxiety Scale (GAI). Results: The findings revealed significant differences between the immigrant and reference samples in TMT time A (p = .018), TMT time B (p = .004), TMT B-A (p = .03 0), GAI (p < .001), and GDS (p = .001), with immigrants performing higher than their peers in all these measures. A Linear regression analysis indicated that having hypertension (p = .026), being unmarried (p = .008), and having fewer years of formal education (p = .003) are associated with higher scores in TMT Time A, while having less formal education (p > .001) is related to higher scores in Time B. Conclusion: The findings suggest that Lusophone immigrants residing in Luxembourg have worse affective and cognitive outcomes compared to their peers living in Portugal. However, further exploration of the influence of health and sociodemographic factors is necessary. [less ▲]

CogPonder: Towards a Computational Framework of General Cognitive Control

Poster (2023, August)

Current computational models of cognitive control exhibit notable limitations. In machine learning, artificial agents are now capable of performing complex tasks but often ignore critical constraints such ... [more ▼]

Current computational models of cognitive control exhibit notable limitations. In machine learning, artificial agents are now capable of performing complex tasks but often ignore critical constraints such as resource limitations and how long it takes for the agent to make decisions and act. Conversely, cognitive control models in psychology are limited in their ability to tackle complex tasks (e.g., play video games) or generalize across a battery of simple cognitive tests. Here we introduce CogPonder, a flexible, differentiable, cognitive control framework that is inspired by the Test-Operate-Test-Exit (TOTE) architecture in psychology and the PonderNet framework in machine learning. CogPonder functionally decouples the act of control from the controlled processes by introducing a controller that acts as a wrapper around any end-to-end deep learning model and decides when to terminate processing and output a response, thus producing both a response and response time. Our experiments show that CogPonder effectively learns from data to generate behavior that closely resembles human responses and response times in two classic cognitive tasks. This work demonstrates the value of this new computational framework and offers promising new research prospects for both psychological and computer sciences. [less ▲]

Graph neural networks for investigating complex diseases: A case study on Parkinson's Disease

Poster (2023, July 25)

Omics data analysis is a critical component in the study of complex diseases, but the high dimension and heterogeneity of the data often pose challenges that are difficult to address by classical ... [more ▼]

Omics data analysis is a critical component in the study of complex diseases, but the high dimension and heterogeneity of the data often pose challenges that are difficult to address by classical statistical and machine learning methods. Recently, structured data analyses using graph neural networks (GNNs) have emerged as a promising complementary approach, particularly for investigating the relational information between samples. However, it is still unclear which strategies for designing and optimizing GNNs are most effective when working with real-world data from complex disorders, such as Parkinson's disease (PD). Our study addresses this gap by examining the application of various GNN models, including Graph Convolutional Network, ChebyNet, and Graph Attention Network, to identify and interpret discriminative patterns between PD patients and controls using omics data. The developed pipeline integrates Lasso penalty-based feature selection, similarity graph construction, and final modeling for sample classification. Through an end-to-end model building and evaluation process, we assess the practical utility of the pipeline on independent PD omics datasets. Overall, our analyses highlight some of the benefits and challenges of using graph structure data for machine learning analysis of disease-related omics data and provide directions for further research. [less ▲]

Education as risk factor of Mild Cognitive Impairment - the role of the gut microbiome

Poster (2023, July)

Novel Geometric Deep Learning Surrogate Framework for Non-Linear Finite Element Simulations

Poster (2023, June 27)

Cognitive Modulation Of Pain By Attention: The Role Of Executive Functioning In Aging

Poster (2023, June 09)

Attention is acknowledged as an effective strategy to modulate pain, as shown by various studies of acute and chronic pain. Here executive functioning (EF) plays a key role, with poorer EF related to less ... [more ▼]

Attention is acknowledged as an effective strategy to modulate pain, as shown by various studies of acute and chronic pain. Here executive functioning (EF) plays a key role, with poorer EF related to less successful inhibition of pain, both relying on the prefrontal cortex (PFC). With age, structural and volume changes in the PFC are accompanied by age-related cognitive decline. So far, only a few studies investigated how age-related cognitive decline affect cognitive distraction from pain.In a two-session design, healthy young (18-30 years) and older participants (60+ years) performed four EF tasks shown to either reveal age-related cognitive decline or not. Afterwards, participants performed a pain distraction task, a n-Back working memory task with low and high cognitive load, during which participants received individually adjusted transdermal electrical pulse trains in non-painful and moderately painful intensities to the inner forearm. Participants rated them regarding their intensity and unpleasantness. Stimulus-related (EF tasks) and pain-related evoked potentials were recorded with a 64-channel EEG.Preliminary results suggest a similar effective subjective pain relief during high and low working memory load in both age groups, which is reflected in a similar reduced N1-P2 pain-related ERPs in older but not in younger participants. However, including executive functioning showing age-related cognitive decline, revealed less successful pain reduction in the high working memory load condition in older adults. These findings could lead to a better understanding how to adapt pain treatments in the older population by including selective cognitive trainings, optimizing pain modulation settings. [less ▲]

The impact of inhibitory control on the acute stress response comparing young and older adults

Poster (2023, June 09)

Background Inhibitory control is a core executive function. It involves our ability to think before we act and allows an individual to control their automatic impulses. Executive functions are a mechanism ... [more ▼]

Background Inhibitory control is a core executive function. It involves our ability to think before we act and allows an individual to control their automatic impulses. Executive functions are a mechanism of the prefrontal cortex, which is highly stress sensitive. Research suggests that executive functions positively influence the stress response, with higher executive functions supporting a more successful stress regulation, but it is unclear if that is also the case in older age. Methods To investigate this, healthy young (18-30 years) and older participants (65+ years) were asked to perform two inhibitory control tasks (Stroop-Color-Word task and Go/Nogo task) and were then assigned to a stress condition (Trier Social Stress Test) and to the control condition in counterbalanced order in a two-session design. Cardiovascular parameters and self-reported stress were used as indices for the stress response. Results First analyses suggest that better inhibitory control is associated with less agitation for the TSST condition in both, young and older adults. While the relationship between response inhibition and the stress response seems to be age-unrelated, preliminary results suggest an age-depended impact of interference control on the physiological stress response. Conclusion Since cognitive decline comes with age, older people may be more affected by stress, this could be counterbalanced by training executive functions. Therefore, these results highlight the importance of inhibitory control and suggest the possibility that enhancing executive function may improve stress management. [less ▲]

Diagnosing specific learning disorder in mathematics in a multilingual education context

Poster (2023, June 07)

Determination of Kinetic Parameters for Heterogeneous Reaction System Employing Discrete Element Methods under HPC Platforms

Poster (2023, June 06)

The complex processes of heterogeneous reactions of granular materials such as occurring during metals-ore reduction or biomass gasification involve numerous physical phenomena. The combination of ... [more ▼]

The complex processes of heterogeneous reactions of granular materials such as occurring during metals-ore reduction or biomass gasification involve numerous physical phenomena. The combination of elevated temperature, complex flow, aggressive atmosphere and heterogeneous chemistry make it difficult to study these industrial processes. One of the most important aspects f heterogeneous reactions is to understand and quantify the evolution of the different transformations. For instance, during metal-oxides reduction processes, it is of high importance to quantify the rate at which the pure metal is formed. Nevertheless, it is almost impossible, by experimental means only, to separately observe, accurately quantify and gain insight into these mingled nonlinear physical and chemical processes. In the last decade, numerical simulation tools for particulate processes, such as the eXtended Discrete Element Method (XDEM), have become indispensable to study complex systems without the need of costly experimental practices. In the past, the XDEM has been employed to predict the reduction of tungsten trioxide (WO 3) in dry hydrogen (H2) atmospheres [1] and reduction of iron ores [2]. In the before-mentioned research works, it was employed kinetic data extracted from literature. On one hand, in these processes the kinetic data differ from each other. This is due to the fact that the experimental data in the literature is interpreted with lumped models and empirical models bonded to the specific experimental conditions. On the other hand, advanced simulation tools, such as XDEM, account for all the influencing phenomena (e.g. species and energy distribution, flow conditions, particles shape, rheological properties) constantly interacting in time and space. In these advanced simulation tools, each particle is treated and solved as individual entities and an accurate prediction of the species formation and transport in time and space is provided. Thus, in such advanced numerical tools, the reaction rate parameters representative of the kinetics alone of the involved chemical reactions must be employed. In this contribution, two XDEM simulation case studies accounting for the industrial reduction of WO 3 are presented. The first case study is employed to determine the reaction rate parameters of the four prevalent reduction steps (WO 3↔WO2.9↔WO2.72↔WO2↔W) upon the H 2 reduction of O3. Where the reaction rates are modeled following an Arrhenius law with two parameters per step i.e. pre-exponential factor and activation energy). The constituted optimization problem of minimization of error of the XDEM simulations vs experimental data, implemented and solved in a High Performance Computing (HPC) cluster, is presented and discussed. The determined parameters are later assessed by comparison to a secondly presented case study. [less ▲]

A High-Resolution Numerical Weather Prediction Model for Nowcasting Precipitation in the Grand-Duchy of Luxembourg (NWPLux)

Poster (2023, June)

Compared to alluvial floods, flash or pluvial floods are difficult to predict because they result from intense and brief periods of extreme precipitation. The project NWPLux aims to develop, a high ... [more ▼]

Compared to alluvial floods, flash or pluvial floods are difficult to predict because they result from intense and brief periods of extreme precipitation. The project NWPLux aims to develop, a high-resolution numerical weather prediction (NWP) model for effective local heavy rainfall prediction in a nowcasting scenario and provide real time for flood simulation. Modelling is based on the Weather Research and Forecasting (WRF) model, with a large-eddy simulation (LES) type 3-dimensional cloud model to simulate small-scale, high-intensity convective precipitation. It is the first such dedicated NWP model for Luxembourg and the Greater Region. As part of this project, we will also test run the LISFlood flood model in an operational inundation forecast mode, meaning that the flood model will be run with the WRF precipitation forecasts as inputs. As an initial run the WRF model was setup with two distinct domains, specifically the Greater Region and Luxembourg with respective horizontal grid spacing of 16 km and 4 km, leveraging high resolution static data. Met data from the NCAR RDA dataset ds083.2 was utilized, spanning from July 08, 2021 to July 15, 2021, and was subjected to the "Conus" physics suit. The output interval was one hour. Preliminary results from NWP runs indicated that the higher resolution simulations (4 km) exhibited superior performance when compared to the lower resolution simulations (16 km). Keywords: NWP, WRF, Flash flood, LISFlood, cloud modelling [less ▲]

Ultra-Light Floating Platform: An Orbital Emulator for Space Applications

Poster (2023, May 30)

Cognitive control of pain in Aging – comparison of different pain modulation strategies

Poster (2023, May 12)

Background. While older people report acute and chronic pain more often than younger people, and, therefore, would benefit significantly from non-pharmacological pain treatment, little is known about how ... [more ▼]

Background. While older people report acute and chronic pain more often than younger people, and, therefore, would benefit significantly from non-pharmacological pain treatment, little is known about how age affects different psychological strategies of pain modulation. The few studies on cognitive distraction from pain suggest a reduced pain relief in older adults, whereas studies on placebo analgesia revealed inconsistent results. So far, distraction and hypnotic analgesia have hardly been investigated in aging. Methods. Healthy young and older participants underwent either a cognitive pain distraction task (working memory task), a placebo analgesia realized with a sham TENS intervention, a hypnotic analgesia intervention or a verbal pain distraction intervention, while receiving non-painful and moderate painful individually adjusted transdermal electrical pulse trains to the inner forearm. Pain ratings and pain-related evoked potentials via 64-channel EEG were recorded. Results. First analyses on the currently small sample suggest a differential impact of age on pain modulation strategies. Since the current sample size is too small to draw reliable conclusions, results will be presented and discussed at the conference. Conclusion. Our results will contribute to a deeper understanding on the efficacy of cognitive pain modulation in aging, helping to optimize pain treatments in this population. [less ▲]

Initial Orbit Determination Results from the University of Luxembourg using Spire GNSS Tracking Data

Poster (2023, April 28)

CubeSats constellations using commercial off-the-shelf components have been studied for different applications, such as GNSS Radio Occultation (GNSS-RO). Furthermore, precise orbit determination of Low ... [more ▼]

CubeSats constellations using commercial off-the-shelf components have been studied for different applications, such as GNSS Radio Occultation (GNSS-RO). Furthermore, precise orbit determination of Low Earth Orbit (LEO) CubeSats based on multiple GNSS constellations would open new opportunities for scientific applications such as Earth’s gravity field measurements. In GNSS kinematic orbit determination, which is the common method used for small sats, the derived orbits are affected by noise, data gaps, outliers, measurement errors as well as poor geometry of the observations. Our work seeks to mitigate these issues and we present two areas of research: 1) GNSS network processing of GPS and Galileo constellations and 2) kinematic orbit determination of a set of Spire CubeSats that host a GNSS-RO payload. An initial architecture of kinematic orbit processing for the Spire GNSS-RO CubeSats constellation is obtained and the details on validations and limitations are discussed in more details. In addition, we showcase the agreement between the GNSS orbit products produced at the University of Luxembourg (UL) with those of the Center for Orbit Determination in Europe (CODE). Finally, the Spire kinematic orbits based on the raw observation approach are derived and compared to the L1B Spire orbit products. [less ▲]

Lernstörung oder Sprachbarriere? – Entwicklung eines für Luxemburg angepassten Wortschatztests

Poster (2023, February 24)

The Ugly Truth of Realistic Perception in Vehicular Simulations

Poster (2023, January 08)

Automated vehicles use sensors to perceive the environment, and studies have shown the limitations of these sensors. The onboard sensors may not detect objects when other participants occlude the Field of ... [more ▼]

Automated vehicles use sensors to perceive the environment, and studies have shown the limitations of these sensors. The onboard sensors may not detect objects when other participants occlude the Field of View (FoV). Thus, sensor efficiency must be tested to ensure its reliability. Simulation is an excellent test option due to the complexity associated with practical experiments. However, emerging simulation frameworks still have various limitations, especially when it comes to large-scale evaluation. This work investigates realistic perception simulation options for autonomous vehicles. We report the perception accuracy for different traffic scenarios. [less ▲]

Comparison of Life Satisfaction between physical activity levels according to age in Luxembourg: Health Behaviour in School-aged Children survey 2022

Poster (2023)

Background: Life satisfaction is a major component of adolescents’ subjective well-being. Physical activity (PA) is associated with the improvement of physical, cognitive, and mental health. Objective ... [more ▼]

Background: Life satisfaction is a major component of adolescents’ subjective well-being. Physical activity (PA) is associated with the improvement of physical, cognitive, and mental health. Objective: This study aimed to analyse adolescents’ life satisfaction (LS) and levels of PA across age in adolescents from Luxembourg. Method: A representative sample of 7089 adolescents (11-to-18-years-old) took part in the Health Behavior in School Aged Children (HBSC) Luxembourg 2022 survey. LS was measured using Cantril Ladder, levels of PA were determined according to the time spend on and frequency of moderate-vigorous and vigorous PA and classified into: Highly active, Active, Low active and Not active. A two-way ANOVA was performed to analyse differences in the LS mean for PA and age groups. Results: All 11-12y have a better LS than the other age groups in all PA groups (p<0.01). However, Highly Active pupils reported better LS than those Not active, Low Active and Active, in all age groups (p<0.01). Low Active (mean=7.26, CI95%7.10-7.43; mean=6.93 CI95%6.78-7.10) reported worse LS than those who were Active in the 13-14y (mean=7.55, CI95%7.44-7.66; p=0.02) and 15-16y (mean=7.29 CI95%7.18-7.40; p<0.01). Conclusion: PA is a factor associated with better LS in adolescents in Luxembourg in all ages. Furthermore, the study showed that although Highly Active pupils have a higher life satisfaction, being physically active is better than Not being active. This study reinforces the importance of PA to maintain a higher life satisfaction through adolescence in Luxembourg. [less ▲]

Comparison of well-being perception between physical activity levels according with age in Luxembourg: Health Behavior in School-aged Children survey 2022

Poster (2023)

Purpose: Youth is a key life phase, as the behaviours established during this phase will mostly continue during adult live. Mental health is a state of well-being (WB) in which individuals realize their ... [more ▼]

Purpose: Youth is a key life phase, as the behaviours established during this phase will mostly continue during adult live. Mental health is a state of well-being (WB) in which individuals realize their own abilities, can cope with the usual stresses of life, can work productively, and is able to contribute to their community. Physical activity (PA) is associated with the improvement of physical, cognitive, and mental health. This study aimed to analyse adolescents’ state of WB in levels of PA across age, by gender in Luxembourg. Methods: A representative sample of 8117 adolescents (11-to-18-years-old) took part in the Health Behaviour in School Aged Children (HBSC) Luxembourg 2022 survey. WB was measured using the WHO-5 Well-being Index, and levels of PA were determined according to the time spend on and frequency of moderate-vigorous and vigorous PA (Highly active, Active, Low active and Not active). A two-way ANOVA stratified by gender was performed to analyse differences of the WB mean for PA levels and age group. Results: For both girls and boys, those Highly active reported better WB than those Not active, in all age groups (p<0.01, all age groups). For the intermediary levels of PA, different gender and age patterns were observed for the WB. While differences were observed between intermediary levels of PA among young girls, no statistically significant difference in mean WB was observed between intermediary levels of PA for girls aged 17 to 18 (highly active (mean=12.9, CI95%11.4-14.5), active (mean=12.4, CI95%11.8-12.9) and Low active (mean=11.80, CI95%11.3-13.2). In boys, only the age group 15-16 showed a difference between the highly active (mean=15.9, CI95%15.2-16.5) and active (mean=15.2, CI95%14.9-15.6) compared to the Low active (mean=13.3, CI95%12.8-13.9) and not active (mean=12.2, CI95%10.6-13.8). Conclusions: Higher PA level is a protective factor for the WB of adolescents in Luxembourg in all ages. This study reinforces the importance of PA to maintain higher levels of well-being through adolescence in Luxembourg. To increase WB in Luxembourg future actions could be taken in order to reinforce the adherence to PA WHO recommendations in Luxembourg in younger ages and promote any level of PA for older ages. [less ▲]

Hybrid Active-Passive Reconfigurable IntelligentSurface-Assisted UAV Communications

Poster (2022, December)

We consider a novel hybrid active-passive reconfigurable intelligent surface (RIS)-assisted unmanned aerial vehicle (UAV) air-ground communications system. Unlike the conventional passive RIS, the hybrid ... [more ▼]

We consider a novel hybrid active-passive reconfigurable intelligent surface (RIS)-assisted unmanned aerial vehicle (UAV) air-ground communications system. Unlike the conventional passive RIS, the hybrid RIS is equipped with a few active elements to not only reflect but also amplify the incident signals for significant performance improvement. Towards a fairness design, our goal is to maximize the minimum rate among users through jointly optimizing the location and power allocation of the UAV and the RIS reflecting/amplifying coefficients. The formulated optimization problem is nonconvex and challenging, which is efficiently solved via block coordinate descend and successive convex approximation. Our numerical results show that a hybrid RIS requires only 4 active elements and a power budget of 0 dBm to achieve an improvement of 52.08% in the minimum rate, while that achieved by a conventional passive RIS with the same total number of elements is only 18.06%. [less ▲]

Early Childhood Education and Care in Luxembourg - Is attendance influenced by immigration background and socioeconomic status?

Poster (2022, November 10)

For decades, researchers have been raising awareness of the issue of educational inequalities in the multilingual Luxemburgish school system. Especially children from families with a migration background ... [more ▼]

For decades, researchers have been raising awareness of the issue of educational inequalities in the multilingual Luxemburgish school system. Especially children from families with a migration background or a lower socio-economic status show large deficits in their language and mathematics competences in comparison to their peers. The same applies to children who do not speak Luxemburgish or German as their first language (Hornung et al., 2021; Sonnleitner et al., 2021). One way to reduce such educational inequalities might be an early and extensive participation in early childhood education and care (ECEC). Indeed, participation in ECEC was found to be positively connected to language and cognitive development in other countries, especially for children from disadvantaged families (Bennett, 2012). However, these children attend ECEC less often (Vandenbroeck & Lazzari, 2014). There are indications that lower parental costs might go hand in hand with a greater attendance of ECEC in general (for a Luxembourgish study, see Bousselin, 2019) and in particular by disadvantaged families (Busse & Gathmann, 2020). The aim of this study is to spotlight the attendance of ECEC in Luxembourg during the implementation of the ECEC reform after 2017 which increased free ECEC hours for all families from 3 to 20 hours a week. We draw on a large dataset of about 35.000 children from the Épreuves Standardisées (ÉpStan, the Luxemburg school monitoring programme) from 2015 to 2021 and investigate which children attend any kind of regulated ECEC service (public, private or family daycare) in which intensity, taking socio-economic and cultural family factors into account. The findings might help to understand in which contexts ECEC attendance should be further encouraged. Implications for future policy decisions are discussed with the goal of further promoting equal educational opportunities for all children. [less ▲]

Do your magic, Harry Potter! popular fiction books and their impact on students’ reading performances

Poster (2022, November 10)

Previous surveys on the reading habits of Luxembourgish secondary school students (conducted within the framework of the national school monitoring programme Épreuves Standardisées (ÉpStan) in 2016 and ... [more ▼]

Previous surveys on the reading habits of Luxembourgish secondary school students (conducted within the framework of the national school monitoring programme Épreuves Standardisées (ÉpStan) in 2016 and 2019) revealed better reading comprehension results in French and German for those students who frequently read printed narrative texts in their leisure time. However, these studies only focused on different reading modes and text types. They did not investigate which digital and/or printed books students actually read for pleasure, nor which text features determine the positive impact narrative texts have on their reading performances: Is it, for example, the language quality, the richness and complexity of content, or simply the amount of written language they need to process? Therefore, we conducted an explorative follow-up survey within the framework of ÉpStan 2020 and asked secondary school students (Grade 7: n=3055; Grade 9: n=5781) to indicate up to three book titles – printed and e-books respectively – they had read in their leisure time. Despite the omnipresence of digital media, preliminary findings show that both age groups prefer paper-based reading activities when reading longer texts (or books) for pleasure. Nevertheless, the most popular text types and book titles are the same for printed books and e-books: Among them, we find the novel series “Harry Potter”, the rather comic-like book series “Gregs Tagebuch”, and the mangas from the “Naruto”/“Boruto” series. The linkage between students’ leisure time reading activities and their ÉpStan reading performances will be drawn, and some first linguistic text analyses of extracts from the most popular book titles will be conducted in order to reveal some of the text features that foster reading comprehension skills. [less ▲]

Integration of backscatter communication with multi-cell NOMA: a spectral efficiency optimization under imperfect SIC

Poster (2022, November 03)

Future wireless networks are expected to connect large-scale low-powered communication devices using the available spectrum resources. Backscatter communications (BC) is an emerging technology towards ... [more ▼]

Future wireless networks are expected to connect large-scale low-powered communication devices using the available spectrum resources. Backscatter communications (BC) is an emerging technology towards battery-free transmission in future wireless networks by leveraging ambient radio frequency (RF) waves that enable communications among wireless devices. Non-orthogonal multiple access (NOMA) has recently drawn significant attention due to its high spectral efficiency. The combination of these two technologies can play an important role in the development of future networks. This paper proposes a new optimization approach to enhance the spectral efficiency of nonorthogonal multiple access (NOMA)-BC network. Our framework simultaneously optimizes the power allocation of base station and reflection coefficient (RC) of the backscatter device in each cell under the assumption of imperfect signal decoding. The problem of spectral efficiency maximization is coupled on power and RC which is challenging to solve. To make this problem tractable, we first decouple it into two subproblems and then apply the decomposition method and Karush-Kuhn-Tucker conditions to obtain the efficient solution. Numerical results show the performance of the proposed NOMA-BC network over the pure NOMA network without BC. [less ▲]

Biomechanische Untersuchung zu verschiedenen Verfahren der Stabilisierung von Insuffizienzfrakturen des vorderen Beckenringes

Poster (2022, October 26)

SMASCH: Facilitating multi-appointment scheduling in longitudinal clinical research studies and care programs

Poster (2022, September 20)

Longitudinal clinical research studies require conducting various assessments over long periods of time. Such assessments comprise numerous stages, requiring different resources defined by ... [more ▼]

Longitudinal clinical research studies require conducting various assessments over long periods of time. Such assessments comprise numerous stages, requiring different resources defined by multidisciplinary research staff and aligned with available infrastructure and equipment, altogether constrained by time. While it is possible to manage the allocation of resources manually, it is complex and error-prone. Efficient multi-appointment scheduling is essential to assist clinical teams, ensuring high participant retention and producing successful clinical studies, directly impacting patient throughput and satisfaction. We present Smart Scheduling (SMASCH) system [1], a web application for multi-appointment scheduling management aiming to reduce times, optimise resources and secure personal identifiable information. SMASCH facilitates clinical research and integrated care programs in Luxembourg, providing features to better manage multi-appointment scheduling problems (MASPs) characteristic of longitudinal clinical research studies and speed up management tasks. It is present in multiple clinical research and integrated care programs in Luxembourg since 2017, including Dementia Prevention Program, the study for Mild Cognitive Impairment and gut microbiome, and the National Centre of Excellence in Research on Parkinson’s disease [2] which encompasses the study for REM sleep behaviour disorder and the Luxembourg Parkinson’s Study. SMASCH is a free and open-source solution available both as a Linux package and Docker image. [less ▲]

Automatic Classification of Peer Review Recommendation

Poster (2022, September 08)

Development of a fair cognitive test battery for multilingual and multicultural children

Poster (2022, August 29)

Poster: Art Installation “Mirror, Mirror” featuring Deepfake and Neural Style Transfer Technologies

Poster (2022, August 15)

The robotic art installation “Mirror, Mirror” is a result of a creative collaboration between an artist and an engineer. The aim of this project is to transform a person’s image and background in order to ... [more ▼]

The robotic art installation “Mirror, Mirror” is a result of a creative collaboration between an artist and an engineer. The aim of this project is to transform a person’s image and background in order to prompt them to examine their version of self and the surrounding reality and make them aware of the developments in the field of social robotics and Artificial Intelligence. [less ▲]

Retrograde procedural memory in Parkinson's Disease and in REM-Sleep Behavior Disorder

Poster (2022, August)

Deep Learning Based Joint Collision Detection and Spreading Factor Allocation in LoRaWAN

Poster (2022, July)

Long-range wide area network (LoRaWAN) is a promising low-power network standard that allows for long-distance wireless communication with great power saving. L oRa is based on pure ALOHA protocol for ... [more ▼]

Long-range wide area network (LoRaWAN) is a promising low-power network standard that allows for long-distance wireless communication with great power saving. L oRa is based on pure ALOHA protocol for channel access, which causes collisions for the transmitted packets. The collisions may occur in two scenarios, namely the intra-spreading factor (intra-SF) and the inter-spreading factor (inter-SF) interference. Consequently, the SFs assignment is a very critical task for the network performance. This paper investigates a smart SFs assignment technique to reduce collisions probability and improve the network performance. In this work, we exploit different architectures of artificial neural networks for detecting collisions and selecting the optimal SF. The results show that the investigated technique achieves a higher prediction accuracy than traditional machine learning algorithms and enhances the energy consumption of the network. [less ▲]

Ergodic Capacity Analysis of NOMA-Based Two-Way Relaying Systems

Poster (2022, July)

In this work, a non-orthogonal multiple access (NOMA) relying on a two-way relaying (TWR) approach is investigated to promote more efficient use of the available spectrum. Specifically, a NOMA-enabled ... [more ▼]

In this work, a non-orthogonal multiple access (NOMA) relying on a two-way relaying (TWR) approach is investigated to promote more efficient use of the available spectrum. Specifically, a NOMA-enabled half-duplex decode-and-forward relay node is used to simultaneously exchange data between two cellular users and a base station (BS), while assuming that the direct links between the two users and the BS are unavailable or negligible. In our analysis, the ergodic capacity is evaluated, in which high signal-to-noise ratio approximations, as well as upper bounds are derived owing to the intractability related to the exact analysis. The results reveal the superiority of the proposed TWR-NOMA scheme compared with the traditional one-way relaying (OWR-NOMA) scheme. Additionally, simulation results show that selecting a relay station (RS) closer to the BS can achieve a higher ergodic sum capacity than selecting a far RS. [less ▲]

Emission-aware Resource Optimization Framework for Backscatter-enabled Uplink NOMA Networks

Poster (2022, June 19)

In the last decade, a sharp surge in the number of user proximity wireless devices (UPWDs) has been observed. This has increased the level of electromagnetic field (EMF) exposure of the users ... [more ▼]

In the last decade, a sharp surge in the number of user proximity wireless devices (UPWDs) has been observed. This has increased the level of electromagnetic field (EMF) exposure of the users substantially and hence, the possible physiological effects. Ambient backscatter communications (ABC) has appeared to be a promising solution to reduce the power consumption of UPWDs by converting ambient radio frequency (RF) signals into useful signals while non-orthogonal multiple access (NOMA) is a compelling multiplexing scheme for enhanced spectral efficiency. This paper utilises a novel combination of ABC and NOMA to reduce the EMF in the uplink of wireless communication systems. This contemporary approach of EMF-aware resource optimization is based on k-medoids and Silhouette analysis. To curtail the uplink EMF, a power allocation strategy is also derived by converting a non-convex problem to a convex one and solving accordingly. The numerical results exhibit that the proposed ABC, NOMA, and unsupervised learning based scheme achieves a reduction in the EMF by at least 75% in comparison to the existing solutions. [less ▲]

Mixed RIS-Relay NOMA-Based RF-UOWC systems

Poster (2022, June)

Reconfigurable intelligent surface (RIS), nonorthogonal multiple access (NOMA), and underwater optical wireless communication (UOWC) are paradigms of technologies that drive the development of ... [more ▼]

Reconfigurable intelligent surface (RIS), nonorthogonal multiple access (NOMA), and underwater optical wireless communication (UOWC) are paradigms of technologies that drive the development of communications nowadays. In this paper, we investigate the performance of a NOMA-based RISassisted hybrid radio frequency (RF)-UOWC system. Due to the interruption of the direct link between the base station and the ship floating on the surface of the water, communication will be carried out via an RIS fixed to an intermediate building. The ship works as a relay that redirects the received signal to two underwater destinations simultaneously. In this paper, we provide new analytical expressions for the outage probability (OP), asymptotic analyses of the OP, and diversity order (D) to gain insights into the system performance. The results showed that the diversity order depends on the UOWC receiver detection technique. In the end, we illustrated that the NOMA-based RISassisted system significantly improves the outage performance of hybrid RF-UWOC systems over a benchmark systems. [less ▲]

CFD-DEM simulation of melt pool formation and evolution in powder bed fusion process

Poster (2022, May 31)

Computational models can be used to optimize metal additive manufacturing parts, and can also play a role in the evaluation of component quality. Among the most important components of such models will be ... [more ▼]

Computational models can be used to optimize metal additive manufacturing parts, and can also play a role in the evaluation of component quality. Among the most important components of such models will be the detailed simulation of flow and heat transfer in and around the melt pool that is formed when the powder bed is melted. In the present work, A Powder Bed Fusion process is studied numerically by using a coupled Computational Fluid Dynamics (CFD) model and eXtended Discrete Element Method (XDEM) model to predict the physical behavior of discrete particles and the melt pool. In XDEM, a randomly packed powder bed of spherical particles is generated and heat and momentum exchange of each particle with other particles and the melt pool are calculated. The CFD model will predict the effects of laser-melt and powder-melt interactions on the melt pool dynamics. Using the developed numerical framework, it will be possible to determine how powder size distribution, the velocity of a laser beam, and the power, among other factors, will affect the characteristics of melt pool. [less ▲]

Exploring NVIDIA Omniverse for Future Space Resources Missions

Poster (2022, May 03)

The resources of space offer a means to enable sustainable exploration of the Moon and Solar System beyond, thus developing space resource technologies is becoming a major topic for space-related activity ... [more ▼]

The resources of space offer a means to enable sustainable exploration of the Moon and Solar System beyond, thus developing space resource technologies is becoming a major topic for space-related activity internationally. However, verifying and validating such systems on Earth conditions is challenging due to the difficulty of filling the sim2reality gap by creating the exact environment. We hypothesize that having on-ground experimental facilities that integrate high-fidelity simulation and physical systems will enable close-to-real testing, speeding up the transition between space technology development and deployment stages. NVIDIA Omniverse recently gained interest to create photorealistic environments, and it is a promising tool to simulate space-related scenarios with high fidelity. Physically accurate and faithful on-orbit scenarios could be generated in Omniverse Create by integrating PhysX physics core and Pixar Universal Scene Description. Omniverse also includes a robotic simulator that connects to physical robotic systems. Various connectors between Omniverse and other platforms such as Unreal Engine, Blender, Autodesk, ParaView, and online collaboration capacity offer the possibility of importing models of space mission components, space scenes, and scientific data into Omniverse. NVIDIA Omniverse seems auspicious in terms of developing high-fidelity photorealistic simulations. In the HELEN project between SpaceR and Spacety, we are developing a close-to-real testing environment for validating debris removal technology. Within this project, we will explore the potential of Omniverse to integrate virtual and physical components, i.e., high-fidelity photorealistic on-orbit simulations with the Zero-G lab facility, for creating reliable testing conditions to reduce the sim2reality gap. SIL and HIL testing architectures for space systems will be developed using software such as MATLAB/Simulink. Moreover, the robotic systems of the Zero-G lab can be linked to the Omniverse’s robotic simulator using its ROS & ROS2 bridge. The figure presents an overview of Omniverse under the scope of the HELEN project. HELEN will show the combination of photorealistic simulations using Omniverse, SIL, and HIL with the Zero-G lab creates a high-fidelity testing environment for future space resources technology. We also believe that the number of human-made objects orbiting the Earth constitutes a great potential for the recovery of their resources. Most of those include valuable materials (Aluminum, Gold, Silver). Therefore, in the future, debris mitigation efforts can target the recovery of such resources, as pointed out in [1]. Reference:[1] Frank Koch, The Value of Space Debris (2021), 8th European Conference on Space Debris [less ▲]

Downlink Throughput of Cell-Free Massive MIMO Systems Assisted by Hybrid Relay-Reflecting Intelligent Surfaces

Poster (2022, May)

We consider in this work a cell-free (CF) massive multiple-input-multiple-output (mMIMO) system where multiple hybrid relay-reflecting intelligent surfaces (HR-RIS) are deployed to assist communication ... [more ▼]

We consider in this work a cell-free (CF) massive multiple-input-multiple-output (mMIMO) system where multiple hybrid relay-reflecting intelligent surfaces (HR-RIS) are deployed to assist communication between access points and users. We first present the signal model and derive the minimum-mean-square-error estimate of the effective channels. We then present a comprehensive analysis for the considered HR-RIS-aided CF mMIMO system, where the closed-form expression of the downlink throughput is derived. The presented analytical results are also valid for conventional CF mMIMO systems, i.e., CF mMIMO systems with and without passive reconfigurable intelligent surfaces. Finally, the analytical derivations are verified by extensive numerical results. [less ▲]

Individual factors and beliefs determining COVID-19 vaccination willingness

Poster (2022, April)

Background: High vaccination coverage rates are necessary to reduce infections and transmissions of the SARS-CoV-2 virus causing COVID-19 and to allow successful mitigation of the current pandemic. To ... [more ▼]

Background: High vaccination coverage rates are necessary to reduce infections and transmissions of the SARS-CoV-2 virus causing COVID-19 and to allow successful mitigation of the current pandemic. To date, we are still lacking information to explain the hesitancy in Luxembourg towards uptake of the available COVID-19 vaccines. The present study explored motivations for and against vaccination in a population-representative sample of residents across Luxembourg to identify hesitant groups and develop strategies to increase population immunity against SARS-CoV-2. Methods: In the framework of the nationwide, representative longitudinal CON-VINCE study, a sample of 1589 respondents (49.6% women, 84.3% Luxembourg nationality) ranging from 18-84 years, participated in the survey in spring 2021. The protocol of the CON-VINCE study has been described in detail elsewhere (Snoeck et al. 2020). Results: 52% of the respondents had at least partial vaccination at time of assessment between April to June 2021. The most common reasons for vaccination of those willing to be vaccinated (81.2%) were altruistic motivations. Prevalent reasons against vaccination for those undecided (8.7%) or reluctant (10.2%) to be vaccinated were that the vaccine had not been tested sufficiently and the fear of long-term vaccine side effects. Only very few of the vaccination-hesitant or -reluctant respondents reported that they did not believe in vaccination in general. Conclusion: The present study identified motivations for and against COVID-19 vaccination and determined demographic and socio-economic factors associated with vaccination willingness. To increase vaccination rates, public health communication needs to target those unsure or unwilling to be vaccinated. We will continue to study the vaccination uptake in the Luxembourg population, as CON-VINCE is now part of the H2020-funded international ORCHESTRA project (https://orchestra-cohort.eu), research into comparing these results on a Pan-European level. [less ▲]

Retrograde Procedural Memory in Parkinson's Disease

Poster (2022, March)

Profiles of cannabis consumption among school-aged children in Luxembourg

Poster (2022)

Background: Cannabis is the most widely used illegal drug globally, with almost 200 million users in 2019. This use is associated with many negative health consequences and social problems, such as low ... [more ▼]

Background: Cannabis is the most widely used illegal drug globally, with almost 200 million users in 2019. This use is associated with many negative health consequences and social problems, such as low school performance and other drug use. Cannabis use can be articularly harmful if it starts early, if it’s excessive or if it occurs associated with other forms of problematic behaviours. A lot of health behaviours, good or bad, emerge during adolescence and remain through the life course. Changing unhealthy behaviours in adolescence would significantly impact society, such as reducing the load of disease, injuries, and economic costs. The objective of this study is to explore cannabis consumption profiles and the factors related to it during adolescence. Methods: This study is based on the 2017/2018 Health Behaviour in School-aged Children Luxembourg survey. Participated 6 880 adolescents, aged 12 to 18-year-old, that attended schools in Luxembourg that teach according to the national. Adolescents answered if they used cannabis during their lifetime and during the past month. Following, they were categorized into 4 groups: 1. never consumed cannabis in a lifetime; 2. consumed in a lifetime but not in the past month; 3. non-daily use during the past month and; 4. daily use (or almost) during the past month. To compare and characterize the groups, descriptive and bivariate analyzes were undertaken, as well as their significance level. Results: Results show that most of the adolescents (81.8%) never consumed cannabis in their lifetime; 9.4% declared a consumption during their lifetime but not the past month; 6.6% used during the past month, non-daily; and 2.2% used (almost) daily during the past month. Different profiles emerge between the groups, particularly with regard to socio-demographic characteristics and risky behaviours. In the case of social relationships and indicators of health and well-being, differences exist but are not as pronounced. Boys, older students and those who do not live with both parents are more likely to be part of groups that have used cannabis. In addition, most risk behaviours show a dose-response relationship with cannabis use. The increase in risky behaviours translates into not only a greater likelihood of using cannabis but also a greater frequency of use. Conclusion: The four different profiles evolve gradually with the frequency of cannabis consumption. This dose-effect relationship has been highlighted between cannabis consumption and the other variables present but is especially true for risk behaviours. Prevention strategies should distinguish adolescents from each of these groups, starting in early adolescence and focus on risk behaviours as a whole. [less ▲]

Why should we discuss gender inequalities in Latin America and their effects in later life?

Poster (2022)

The influence of tens and units inversion in a two-digit number matching task in German and French monolinguals and bilinguals

Poster (2022)

The pandemic as a stress test of the patent system – a legal-economic re-examination of exclusivity, liability rules, open innovation, and complementary policy levers

Poster (2021, November 15)

Post-Urban Mining Automation and Digitalisation for a Closed- Loop Circular Construction

Poster (2021, October 13)

The large volume of in- and out-flow of raw materials to construction projects has a huge potential to be optimised for resource efficiency and waste reduction. With the recent awareness of the importance ... [more ▼]

The large volume of in- and out-flow of raw materials to construction projects has a huge potential to be optimised for resource efficiency and waste reduction. With the recent awareness of the importance of the circular economy, construction actors are aligning their practices to be more circular and sustainable. The concept of material banks is born out of this awareness in order to document the lifecycle information of materials and facilitate re-using them. The introduction of new cycles before individual materials reach their final lifecycle stages results in reduced negative environmental impacts. This paper presents a workflow by positioning different digital technologies to automate the procedures for reuse assessment: from the deconstructed building to M/C bank to new construction projects. This automation supports a practical material and component reuse, while it provides the necessary infrastructure to digitise and digitalise the post-deconstruction materials to be visualised, selected and used by future designers in Building Information Modelling (BIM)-based design and management environments. To this aim, the coupling of BIM, reality capturing technologies, additive manufacturing techniques, IoT and RFID sensors is also anticipated. [less ▲]

Massive MIMO Downlink Transmission for LEO Satellite Communications

Poster (2021, September)

We investigate the downlink (DL) transmit strategy for massive multiple-input multiple-output (MIMO) low-earthorbit (LEO) satellite communication (SATCOM) systems, in which only the slow-varying ... [more ▼]

We investigate the downlink (DL) transmit strategy for massive multiple-input multiple-output (MIMO) low-earthorbit (LEO) satellite communication (SATCOM) systems, in which only the slow-varying statistical channel state information is known at the transmitter side. First, we establish the massive MIMO LEO satellite channel model, in which the uniform planar arrays are deployed at both the satellite and user terminals (UTs). Building on the rank-one property of satellite channel matrices, we show that transmitting a single data stream to each UT is optimal for the ergodic sum rate maximization. This result is of great importance for massive MIMO LEO SATCOM systems, since the sophisticated design of transmit covariance matrices is turned into that of precoding vectors, with no loss of optimality at all. Furthermore, we conceive an algorithm to compute the precoding vectors. Simulation results show the significant performance gains of the proposed approaches over the previous schemes. [less ▲]

Developing multiliteracies in early years education in Luxembourg

Poster (2021, August 17)

Model bias and its impact on computer-aided diagnosis: A data-centric approach

Poster (2021, August)

Machine learning and data-driven solutions open exciting opportunities in many disciplines including healthcare. The recent transition to this technology into real clinical settings brings new challenges ... [more ▼]

Machine learning and data-driven solutions open exciting opportunities in many disciplines including healthcare. The recent transition to this technology into real clinical settings brings new challenges. Such problems derive from several factors, including their dataset origin, composition and description, hampering their fairness and secure application. Considering the potential impact of incorrect predictions in applied-ML healthcare research is urgent. Undetected bias induced by inappropriate use of datasets and improper consideration of confounders prevents the translation of prediction models into clinical practice. Therefore, in this work, the use of available systematic tools to assess the risk of bias in models is employed as the first step to explore robust solutions for better dataset choice, dataset merge and design of the training and validation step during the ML development pipeline. [less ▲]

How to categorize job demands? Revising the challenge-hindrance distinction of job demands

Poster (2021, July 19)

𝑘-Pareto Optimality for Many-Objective Genetic Optimization

Poster (2021, July)

Unlinkable Updatable Hiding Databases and Privacy Preserving Loyalty Programs

Poster (2021, July)

Loyalty programs allow vendors to profile buyers based on their purchase histories, which can reveal privacy sensitive information. Existing privacy friendly loyalty programs force buyers to choose ... [more ▼]

Loyalty programs allow vendors to profile buyers based on their purchase histories, which can reveal privacy sensitive information. Existing privacy friendly loyalty programs force buyers to choose whether their purchases are linkable. Moreover, vendors receive more purchase data than required for the sake of profiling. We propose a privacy-preserving loyalty program where purchases are always unlinkable, yet a vendor can profile a buyer based on her purchase history, which remains hidden from the vendor. Our protocol is based on a new building block, an unlinkable updatable hiding database (HD), which we define and construct. HD allows the vendor to initialize and update databases stored by buyers that contain their purchase histories and their accumulated loyalty points. Updates are unlinkable and, at each update, the database is hidden from the vendor. Buyers can neither modify the database nor use old versions of it. Our construction for HD is practical for large databases. [less ▲]

Investigating the associations of trajectories of depressive symptoms and self-perceived health and incident dementia : an unsupervised machine learning approach.

Poster (2021, July)

Background: Risk factors for dementia show inter-individually varying trajectories over the lifespan. However, risk factors have been mainly investigated with one time-point assessments. New research ... [more ▼]

Background: Risk factors for dementia show inter-individually varying trajectories over the lifespan. However, risk factors have been mainly investigated with one time-point assessments. New research suggests that certain risk factor trajectories are associated with increased risk of adverse cognitive outcomes (Demnitz et al., 2020 [https://doi.org/10.1101/2020.05.20.20106963]; Singh-Manoux et al., 2018 [https://doi.org/10.1016/j.jalz.2017.06.2637]). However, it remains unclear how sequential and simultaneous changes of risk factors alter the individual risk for developing dementia. Testing the joint contribution of trajectories of depressive symptoms and self-perceived health on incidence of dementia, we hypothesized that consistently poor as well as deteriorating trajectories increase the risk for incident dementia, and explored possible interactions of the trajectories. Method: A total of 5,326 respondents to the SHARE survey, mean age 73.9 years, and 6 complete follow-ups spanning ~13 years, answered the EURO-D depression scale, self-perceived health (SPH) (t1-t5), and self-reported dementia diagnosis at last follow-up (t6). To investigate the predictive ability of distinct longitudinal trajectories, we applied unsupervised statistical learning methods (K-means cluster modelling). Clusters indicated distinct risk factor trajectories, which were used as exposures in stepwise logistic regressions to predict incident dementia, controlling for age, gender, education, and country. Result: Cluster analysis revealed five distinct trajectories each for SPH and EURO-D, with varying dementia incidence. In stepwise logistic regressions, respondents with trajectories “consistently poor health” and “consistently high depression” showed elevated risk of dementia (OR = 4.02 [1.39, 14.75] and OR = 2.26 [1.03, 4.95], respectively) compared to the combined baseline risk for “consistently low depression” and “consistently good health”. Interactions were not significant. However, respondents with the combination of “consistently high depression” and “consistently poor health” showed increased risk (N = 246; 6.1% dementia). Conclusion: Applying unsupervised machine learning is helpful to incorporate longitudinal information on depressive symptoms and self-perceived health and model these risk factors longitudinally to test their contribution to explain incidence of dementia. The predictive ability of the trajectories of depressive symptoms and self-perceived health for dementia indicates the potential for improving the identification of people at risk for developing dementia in late life by exploiting trajectory information readily accessible through regular medical check-ups in old age. [less ▲]

Does Physical Activity Moderate The Influence Of Sedentary Behavior On Health In Young People?

Poster (2021, June)