International Migration, Integration and Social Cohesion online publications
Not a member yet
484620 research outputs found
Sort by
Democratie en vertegenwoordiging van het algemeen belang:De kabinetsplannen voor ontvankelijkheid van NGO’s in het licht van <i>KlimaSeniorinnen</i>
No full textIn 2023 heeft de Tweede Kamer een motie aangenomen waarin zij de regering verzocht de representativiteit van belangenorganisaties met een ideëel doel overeenkomstig artikel 3:305a Burgerlijk Wetboek (BW) te onderzoeken en te verkennen in hoeverre hier nadere eisen aan gesteld moeten worden. Dit artikel betoogt dat het KlimaSeniorinnen-arrest van het Europees Hof voor de Rechten van de Mens (EHRM) niet alleen onderstreept hoe belangrijk de vertegenwoordiging van klimaatbelangen in een moderne complexe samenleving is, maar ook duidelijke grenzen stelt aan de mate waarin de Nederlandse wetgever de representativiteitsvereiste kan aanscherpen voor klimaatzaken
Data augmentation for vehicle detection with diffusion-based object inpainting
No full textAutomated vehicle detection in video footage captured by Unmanned Aerial Vehicles (UAVs) is a critical capability in security and defense domains, especially for environments where communication is jammed. Development of deep learning-based object detectors for this purpose typically requires large-scale datasets, which can be hard to obtain due to limited access to relevant environments. To address this challenge, synthetic data has been proposed as a supplementary source of training data, introducing additional variations in the appearance and positioning of objects. One promising strategy for generating synthetic data is inpainting, where objects of interest are seamlessly integrated into various backgrounds. However, traditional inpainting techniques lack spatial and contextual awareness, limiting their effectiveness for data augmentation. Recent advancements in generative AI, specifically diffusion models, have demonstrated improvements in object harmonization and spatial control for object inpainting, enabling realistic foreground-background matching with a high level of diversity. In this work, we explore the value of diffusion-based inpainting as a data augmentation technique. We use the inpainting model AnyDoor to enrich a small subset (1000 frames), of the VisDrone train dataset with inpainted versions of minority-class objects (buses, vans, trucks). We train YOLOX detectors on datasets with increasing amounts of synthetic vehicles (1x, 5x, 10x, and 20x) and analyze the impact on detection performance. Results show that zero-shot inpainting can substantially improve detection for buses up to an augmentation factor of 10x, with no improvements at 20x. Effects for vans and trucks are mixed and sometimes negative. Fine-tuning AnyDoor provided limited additional benefit under the tested conditions. Overall, diffusion-based inpainting shows potential as a data augmentation strategy in low-resource UAV scenarios. Future work should explore strategies to increase contextual diversity, such as adding multiple synthetic objects per image or incorporating automated quality control for synthetic samples
Evaluating Locally Run Large Language Models on Toxic Meme Analysis
No full textToxic memes easily spread online, propagating stereotypes, hate, and other stronger or more nuanced types of malicious content. The sheer volume of memes requiring moderation calls for automated methods, but their multiple layers of meaning make them challenging to assess: in some cases, toxicity may stem from subtle wordplay, in others by visual references or evoking hateful symbols, etc. Large language models (LLMs) offer a promising tool for performing toxicity detection, since they can leverage a large amount of contextual information and analyzing content items in depth. In this paper, we investigate the suitability of locally run LLMs to perform such a task. Locally run large language models have several advantages over web-based models like OpenAI’s ChatGPT with respect to costs, reproducibility, and data safety. We evaluate the local models on the tasks of automatic meme analysis and toxic symbol identification, and compare the results with analyses of the online model ChatGPT. Our findings reveal that while local models identify only a limited number of toxic memes and symbols, their labels are often accurate (low recall, high precision). Although they do not achieve perfect performance, we believe these models can effectively support human content moderators.</p
From Content-Based Image and Video Retrieval With Relevance Feedback to Multimodal Large Language Models and Beyond: A Journey Through Modalities and Semantic Levels
No full textThe multimedia community has made a long journey from early content-based image and video retrieval systems, which often facilitated interactive search, exploration, and learning using relevance feedback and similar techniques, to modern foundation models, such as multimodal large language models (MLLMs). During that period, multimedia analysis techniques have improved dramatically, maturing from the low semantic levels of colours, textures, and shapes, over semantic concepts, actions, and events to increasingly abstract and complex multimedia content understanding. With the advent of foundation models, the traditional search and recommendation paradigms have been increasingly replaced with multimodal multi-turn conversational search. In this keynote, I will reflect on this journey by showcasing representative works, including those by our team, advocating that the trends surrounding modern MLLMs have been a product of gradual evolution, rather than a revolution. In addition, I will illustrate how MLLMs can be applied to advance business and society by injecting domain knowledge from the other disciplines and facilitating improved interaction with the user. Finally, I will share some thoughts on the potential of recent research developments aiming to improve performance while significantly reducing the size, complexity and cost of MLLMs
Political socialization across places:differential effects on multiculturalist attitudes in urban and rural areas?
No full textExtensive literature shows that inhabitants from urbanized areas tend to have more multicultural attitudes than those from rural areas, and these differences seem especially pronounced among younger generations. To explore why, we study ‘the formative years of adolescence’ and focus on an often-overlooked mechanism in the literature: political socialization through discussing politics with various actors such as parents, peers, and teachers. We argue that political socialization differently affects multicultural attitudes in urban and rural areas, and test our hypotheses by using the Dutch Adolescent Panel on Democratic Values (DAPDV) dataset, a representative panel study among adolescents in the Netherlands between ages 11 and 16. This allows us to not only analyze whether the associations between political discussion and multicultural attitudes differ between urban and rural areas, but to also disentangle influence from selection effects. We find that political socialization through parents and peers is related to multicultural attitudes differently in urban and rural contexts. It is associated with more multiculturalism in urban areas but not in rural areas. These associations are fully the result of selection mechanisms, and no support for the direct influence of political socialization on multicultural attitudes is found. Overall, our findings stress the importance of further studying the role of political socialization in adolescence to help us understand increasing geographic political polarization across many democratic societies
Deep learning-based derivation of physiological information from cardiac CT angiography
Get PDFCoronary artery disease (CAD) is a leading cause of death worldwide. It is characterized by the accumulation of atherosclerotic plaque in the coronary arteries. This buildup can lead to stenosis, a narrowing of the arteries, which is functionally significant if it results in myocardial ischemia. The current clinical standard for determining the functional significance of stenosis is given by invasive fractional flow reserve (FFR) measurement. However, this procedure is invasive, costly and burdensome for patients.While coronary CT angiography (CCTA) allows for the visual identification of most functionally significant stenoses, it suffers from low specificity. This causes a significant number of unnecessary invasive FFR measurements, emphasizing the need for improved patient selection schemes.This thesis addresses the non-invasive selection of patients who may require invasive treatment by using machine learning (ML), specifically deep learning, to derive FFR from CCTA. Deep learning enables the automatic extraction of robust features from high-dimensional input data, which facilitates accurate and fast FFR prediction without the need for manual intervention. The ultimate goal of this work is to enable automatic and non-invasive FFR prediction, providing clinicians with a tool that can assist in assessing coronary artery disease and making informed decisions about whether invasive procedures are necessary. To achieve this goal, several steps were taken, including the design of a multi-step deep learning model that localizes and characterized the arteries into supervised and unsupervised features, to balance model interpretability and performance
Improving outcomes in minimally invasive pancreatoduodenectomy
Get PDFThis thesis investigates various aspects of minimally invasive pancreatoduodenectomy (MIPD), focusing on its implementation, clinical outcomes, and comparative effectiveness within the Netherlands and across Europe. It highlights the ongoing shift from laparoscopic pancreatoduodenectomy (LPD) to robotic-assisted pancreatoduodenectomy (RPD), supported by data from both national and international registries. The European registry for Minimally Invasive Pancreatic Surgery (E-MIPS) and the randomized controlled DIPLOMA-2 trial form the backbone of this work. The DIPLOMA-2 trial, a multicenter randomized controlled study, demonstrates that MIPD is non-inferior to open pancreatoduodenectomy (OPD) in terms of postoperative complication rates and provides superior outcomes in time to functional recovery.Further analyses confirm that MIPD is both feasible and safe across a broad range of patient populations, including elderly patients, with comparable mortality outcomes to those following OPD. However, inter-institutional variation in outcomes remains. Additionally, the thesis addresses postoperative risk stratification by evaluating fistula risk scores specific to RPD, aiming to enhance perioperative planning and optimize patient selection.Collectively, the findings of this thesis support the safe and evidence-based expansion of MIPD. They underscore the importance of ongoing evaluation through randomized controlled trials and high-quality clinical registries to ensure standardized, safe, and effective surgical practice
Psychopathology and violent extremism:The need for primary data and assessment of countering violent extremism (CVE) policies
Get PDFThis dissertation aims to contribute to our knowledge on the prevalence and relevance of psychopathology in violent extremist samples. It also adds empirical research on whether gender and psychopathology correlate to higher levels of violent extremism in the case of radicalising individuals. The latter information is incorporated in research on Countering Violent Extremism (CVE) policies to better understand challenges concerning psychopathology. As a result, this work contributes to a more balanced approach to the goal of lowering the chances of mentally ill persons committing violent extremist actions without harming and stigmatising the vast majority of mentally ill non-violent individuals (with extremist ideas). The findings in this dissertation show that psychopathology cannot be used as a predictor for violent extremism in the general population. In very rare cases in the general population, mental disorders could be relevant for highly individualised pathways which, combined with other factors, may ultimately culminate in violent extremist activity. However, once mental disorders are present, their roles may vary or be irrelevant. Additionally, although individualised gender-responsive approaches seem needed, no evidence is found that psychopathological profiles of radicalising women differ from their male counterparts. Finally, this dissertation displays how heterogeneous psychopathological roles could complicate the CVE-assessment of (potential) mentally ill violent extremists. This complexity runs through practitioners’ challenges in CVE about case-inclusion, case-management and case-outflow. Some Dutch practitioners, for instance, express concerns on whether they can ethically include severely mentally ill individuals in CVE-programmes. These conclusions act as springboards for renewed consultation within CVE-approaches and research agendas
Traumatic foundations and intersecting paths:Change mechanisms and treatment of childhood trauma, childhood-related posttraumatic stress disorder, and borderline personality disorder
Get PDFWe studied several interconnected concepts: childhood trauma (CT), posttraumatic stress disorder (PTSD), and borderline personality disorder (BPD). When we examined the link between five individual CT subtypes (i.e., sexual, physical, and emotional abuse, and physical and emotional neglect) and PTSD severity, only higher exposure to childhood emotional abuse was related to a higher PTSD severity. Additionally, we studied two PTSD treatments: eye-movement desensitization and reprocessing (EMDR) and imagery rescripting (ImRs). We examined the role of non-fear emotions (i.e., shame, guilt, anger, and disgust) in PTSD changes during treatment and found several differences between changes in these emotions and changes in PTSD symptoms in the short and the long term. We also directly compared the change mechanisms of EMDR and ImRs. The findings indicated that ImRs works via changes in the distress and encapsulated belief ratings related to the trauma memories. In contrast, our findings did not support the hypothesis that EMDR works via changes in memory vividness. Moreover, we studied whether CT and dissociation were moderators of the effectiveness of different formats of schema therapy (ST) for BPD. We found that dissociation was associated with higher treatment retention for a combined individual and group ST format, compared to group ST and treatment-as-usual. Lastly, we examined the effectiveness of psychological treatments for BPD in a multilevel meta-analysis. We analyzed within-treatment effects, and we included a wide range of study designs. The findings indicated that several specialized treatments are related to higher effect sizes compared to the average of all treatments
The unexplored desert of Enterovirus C and the search for antivirals:Insights from human organotypic models
Get PDFEnteroviruses (EVs) are ubiquitous pathogens causing a spectrum of symptoms in humans, ranging from mild to severe. While much attention is given to well-known genotypes like poliovirus, understanding the less-studied non-polio genotypes within the Enterovirus C (EV-C) species is essential for unravelling their molecular epidemiology, tropism, and pathogenesis. Despite predominantly affecting African children, the occurrence of severe cases among immunocompromised patients in Europe underscores the necessity for deeper comprehension. Utilizing human-based organotypic models of the airway, gut, and brain provides promising tools for studying host-pathogen interactions and conducting preclinical investigations with enhanced translational potential. The thesis begins with a global overview of EV prevalence, elucidating genotype distribution variations. EV-C tropism is then explored using organotypic cultures to delineate replication dynamics and infection patterns across respiratory and intestinal models. A clinical case of severe chronic diarrhea in an immunocompromised patient highlights the urgency for effective treatment strategies. The latter part of the thesis explores the utility of diverse organoid models for antiviral testing. A novel method for generating airway organoids with reversed polarity is introduced. Furthermore, we assess the broad-spectrum antiviral efficacy of Halofuginone Hydrobromide against various viruses, including EV-A71, using human organoid models. Finally, discussions on the potential of human-based organotypic models in EV-C research and antiviral testing, drawing parallels with poliovirus animal research, underscore the importance of standardized models and cautious result interpretation. This thesis provides valuable insights into EV-C pathogenesis and emphasizes the significance of innovative and relevant methodologies in testing treatment efficacy against viral diseases