742875 research outputs found
Sort by
One size fits all? Transferring social mindfulness measures to HRI
Applying psychological measures to Human Robot Interaction (HRI) has become increasingly common. Among these, the Social Mindfulness Paradigm (SoMi)has been used to study social mindfulness towards robots through online experiments using vignettes. This line of work indicated that humans do not show prosocial behavior towards robots. However, these findings are potentially confounded in two ways: items in the SoMi task were based on human-human interactions (HHI), not HRI, and experiments did not involve real-life interactions with robots. Addressing these methodological shortcomings, the current studies investigated whether the SoMi task is a valid assessment of social mindfulness in HRI to determine under which conditions, if any, we observe prosocial behavior towards robots. In Study One, participants interacted with a social robot (Cozmo) for three days, with perceived anthropomorphism and social mindfulness assessed before and after the interaction period. In Study Two, participants played the classic version of the SoMi paradigm using revised items matched in value for humans and robots, based on prior evaluations by a separate sample. Prolonged interaction with Cozmo did not increase social mindfulness but increased anthropomorphic perception of the robot. The revised items did not increase social mindfulness in the anthropomorphic condition, but they increased overall social mindfulness compared with previous studies. We conclude that real-life interaction does not necessarily enhance social mindfulness towards robots, that the item selection and their value for both human and robots must be considered, and that future studies should explore other interaction time frames and items.. Further, the increase in perceived anthropomorphism after a period of real-life interaction supports theory on anthropomorphism as a dynamic process. More general, the results stress that the field should carefully test HHI measures to ensure measurement validity before transferring them to HRI and that researchers must consider the context in which HRI occurs for external validity. Our findings also raise new questions for theory on social mindfulness, and support the emerging critique of the widely used Computers Are Social Actors (CASA) theory, which lead to the emergence of psychological measures in the field of HRI.</p
Building Molecules by a Self-Replicator That Catalyzes Acyl Hydrazone Formation
Catalysis of bond-forming reactions is key to the development of life-like chemical systems as it allows to build up new material, increasing molecular complexity and diversity. Integrating catalysis with other characteristic properties of life, like self-replication, represents an important advance in the transition from chemistry to life. We have previously shown that catalysis can emerge in synthetic self-replicators that form through supramolecular assembly. However, the organocatalyzed reactions were solely bond-breaking so far. We now report the successful expansion of the catalytic promiscuity of these systems to bond-forming reactions. We show that a self-replicator efficiently catalyzes acyl hydrazone formation between different hydrazides and aldehydes. This marks an important step towards the further development of evolvable systems that combine metabolic activity with self-replication.</p
Improving medication safety in patients on antithrombotic therapy
Every hour, 11 people in the Netherlands develop thrombosis, a blood clot in a blood vessel. Blood thinners are used to prevent and treat thrombosis, but they can also cause bleeding. Their use requires a careful balance between the risk of bleeding and the risk of thrombosis.After thrombosis, some patients use two blood thinners at the same time. After a few weeks or months, one of these blood thinners should be stopped. In daily practice, this does not always happen.The aim of this thesis is to explore how blood thinners can be used more safely.In a study in three hospitals, we found that 41.2% of patients using two blood thinners should have stopped one blood thinner earlier. Because they did not stop the blood thinner in time, they unnecessarily had a higher risk of bleeding. After advice from a pharmacist, this percentage dropped to 2.2%.Community pharmacists need the right information to prevent patients from using two blood thinners for too long, but this information is often missing. Therefore, we developed a questionnaire that helps pharmacists ask the right questions to patients to decide if using two blood thinners is really needed.We also sent pharmaceutical discharge letters to community pharmacies. These letters explained why patients must use blood thinners and for how long. As a result, patients stopped the blood thinner that needed to be stopped more often on time. The percentage of patients using two blood thinners for too long decreased from 22.1% to 9.3%
Sonic-o1:A Real-World Benchmark for Evaluating Multimodal Large Language Models on Audio-Video Understanding
Multimodal Large Language Models (MLLMs) are a major focus of recent AI research. However, most prior work focuses on static image understanding, while their ability to process sequential audio-video data remains underexplored. This gap highlights the need for a high-quality benchmark to systematically evaluate MLLM performance in a real-world setting. We introduce SONIC-O1, a comprehensive, fully human-verified benchmark spanning 13 real-world conversational domains with 4,958 annotations and demographic metadata. SONIC-O1 evaluates MLLMs on key tasks, including open-ended summarization, multiple-choice question (MCQ) answering, and temporal localization with supporting rationales (reasoning). Experiments on closed- and open-source models reveal limitations. While the performance gap in MCQ accuracy between two model families is relatively small, we observe a substantial 22.6% performance difference in temporal localization between the best performing closed-source and open-source models. Performance further degrades across demographic groups, indicating persistent disparities in model behavior. Overall, SONIC-O1 provides an open evaluation suite for temporally grounded and socially robust multimodal understanding. We release SONIC-O1 for reproducibility and research: Project page: https://vectorinstitute.github.io/sonic-o1/ Dataset: https://huggingface.co/datasets/vector-institute/sonic-o1 Github: https://github.com/vectorinstitute/sonic-o1 Leaderboard: https://huggingface.co/spaces/vector-institute/sonic-o1-leaderboar
Real World Lung Cancer Screening Challenges:Lung nodule management and integration of AI
Lung cancer screening has the potential to save millions of lives. Sufficient evidence has now been provided to support the implementation of lung cancer screening worldwide. Implementation, however, is complex and presents new challenges which mandate further scientific research. Key findings from this thesis could be used to help address some of the challenges associated with the implementation of a lung cancer screening program. Part I (chapter 1 and 2) explored existing literature and international expert opinions to comprehensively review and highlight challenges associated with the implementation of global lung cancer screening programs. Part II (chapters 3 and 4) investigated the association between Influenza season and incidence of baseline detected lung nodules using data collected during the ImaLife general population study, and also resolution of new nodules detected in the NELSON study which included participants at high-risk of developing lung cancer. Part III (chapters 5-7) introduced the concept of using AI as a first-reader to rule out negative baseline CT scans in order to achieve significant reduction in radiologist CT-reading workload. This concept was explored using data from MLCS and then validated in the UKLS dataset with histological outcomes providing the gold standard. Finally, supporting evidence on the implementation of AI as a first reader was provided using data from the ongoing 4ITLR study. In this chapter, a summary of main findings per chapter will be presented followed by general discussion, future perspectives and conclusions
Rate control in atrial fibrillation:A focus on calcium channel blockers
This thesis demonstrates that high ventricular heart rates during AF are common, particularly in patients with paroxysmal atrial fibrillation (AF). Continuous heart rate monitoring with implantable loop recorders revealed that many patients paroxysmal AF spent a large proportion of AF episodes with heart rates above the guideline-recommended threshold of 110 beats per minute. Importantly, rate-control therapy was adjusted in only a minority of these patients, highlighting the practical challenges of achieving adequate rate control, especially due to the risk of bradycardia during sinus rhythm.In patients with newly diagnosed AF, more than one third presented with heart rates above 110 beats per minute at baseline. These patients reported more severe AF-related symptoms. Early rhythm control did not significantly reduce symptoms compared with usual care, suggesting that heart rate during AF is a key determinant of symptom burden.Meta analyses of rate-control medications showed that non-dihydropyridine calcium channel blockers and beta-blockers are similarly effective in lowering heart rate during AF. In retrospective analyses preformed in this thesis calcium channel blockers were associated with a significantly lower incidence of bradycardia during sinus rhythm. Furthermore, patients treated with calcium channel blockers showed signals of reduced AF progression, better exercise tolerance, and higher maximal heart rates during physical activity.Overall, this thesis demonstrates that rate control is often suboptimal in paroxysmal AF and that the choice of rate-control medication may have clinically meaningful implications, supporting a more individualized approach to AF management
Circular citizenship behaviours to promote systemic change:Influences of values, beliefs, norms, and personal agency
Environmental problems arise from our current societal and economic systems and could be alleviated by transforming such systems towards more sustainability. Citizens can engage in behaviours promoting such systemic changes. Based on the Systemic Change through Citizen Action framework, we examine different types of circular citizenship behaviours (CCBs) that reflect actions citizens can take to influence other citizens, governments, and businesses to instigate societal change towards sustainability, and more specifically, a circular economy. Specifically, we aim to study to what extent people engage in these actions and which factors increase the likelihood of engagement. A cross-sectional survey with a representative Dutch sample shows that people rarely engage in CCBs aimed at other citizens, and very rarely in CCBs aimed at governments or businesses. Our findings further indicate that an extended value-belief-norm theory is successful in explaining engagement in CCBs, especially CCBs aimed at other citizens. Besides stronger personal norms, stronger biospheric values and higher outcome efficacy also directly relate to more engagement in most CCBs, while stronger hedonic values relate to less engagement. Interestingly, egoistic values are positively related to CCBs, suggesting that CCBs have different qualities from many other pro-environmental behaviours. Our findings highlight much untapped potential for systemic change through citizen action and offer insights into how engagement in CCBs might be promoted.</p
A health-oriented analytical reasoning approach in a child that fails to thrive
An 11-month-old boy growing up in a vulnerable family environment is not thriving well. The youth healthcare physician provides nutritional advice and refers to a pediatrician. The pediatrician rules out physical causes, gives further nutritional advice, and refers to a dietitian, social worker, and eventually back to youth health care. We wonder whether this referral spiral and unnecessary healthcare costs could have been avoided by using a health-oriented analytical reasoning approach. The method of Health-oriented Analytical Reasoning results in a broad diagnosis that includes contextual factors and appropriate interventions. In this case, we would have focused on building trust with the family, improving nutrition, and reducing stress. There was sufficient reason not to refer to the pediatrician, as no clear signs of physical illness were present. Furthermore, this case highlights the importance of strong collaboration in child health care.</p
Spider Silk Inspired Processing of Liquid Crystalline Complex Coacervates
Spider silk spinning begins with coacervation into a dense protein phase that organizes into liquid crystalline domains. Changes in salt concentration, together with shear forces, then direct the alignment needed to form highly ordered fibers. Inspired by this process, we developed a fully synthetic system of liquid crystalline complex coacervates designed to replicate the hierarchical organization and alignment mechanisms of spider silk, focusing on processing pathways. We show that salt concentration (tetrabutylammonium bromide, TBAB) governs the balance between isotropic and liquid crystalline states, with coacervation suppressed above 0.5 M, smectic order stabilized at ≤0.2 M, and isotropic chain networks prevailing at intermediate concentrations. Crucially, the degree of shear alignment depends strongly on salt: higher salt concentrations accelerate molecular relaxation and raise the threshold shear rate required to induce ordering, echoing the cooperative role of the ion composition and shear in natural silk spinning. Rheological and X-ray scattering measurements confirm that this salt–shear interplay dictates both the viscoelastic response and the molecular anisotropy. Finally, we demonstrate directional alignment through stretching and extrusion-based 3D printing and show that the unique tunability of salt concentration provides direct control over both processability and shear-induced alignment, offering a powerful biomimetic route to anisotropic material design.</p