154092 research outputs found
Sort by
Werkervaringen van zorgmedewerkers met AI:een kwalitatieve verkenning vanuit het AMO-model
Get PDFDe zorgsector wordt in toenemende mate geconfronteerd met uitdagingen als gevolg van groeiende vraag (o.a. door vergrijzing en complexiteit van zorg) en afnemend aanbod van zorgverleners (o.a. door personeelstekorten). Kunstmatige Intelligentie (AI) wordt als mogelijke oplossing gezien, maar wordt vaak vanuit een technologisch perspectief benaderd. Dit artikel kiest een mensgerichte benadering en bestudeert hoe zorgmedewerkers het werken met AI ervaren. Dit is belangrijk omdat zij uiteindelijk met deze applicaties moeten werken om de uitdagingen in de zorg het hoofd te bieden. Op basis van 21 semigestructureerde interviews met zorgmedewerkers die AI hebben gebruikt, beschrijven we de werkervaringen met AI. Met behulp van het AMO-raamwerk - wat staat voor abilities, motivation en opportunities - laten we zien dat AI een impact heeft op het werk van zorgmedewerkers. Het gebruik van AI vereist nieuwe competenties en de overtuiging dat AI de zorg kan verbeteren. Daarbij is er een noodzaak voor voldoende beschikbaarheid van training en ondersteuning. Tenslotte bediscussiëren we de implicaties voor theorie en geven we aanbevelingen voor HR-professionals
Practical Biometric Search under Encryption:Meeting the NIST Runtime Requirement without Loss of Accuracy
Get PDFBiometric search consists of comparing a biometric probe with biometric references in a database to find identities that match the probe. In addition to being resource-intensive, biometric search can lead to privacy violations when biometric data is exposed. To protect the sensitive biometric data, state-of-the-art search solutions are based on fully homomorphic encryption (FHE). However, they either reduce search accuracy to gain efficiency or lack efficiency and fail to meet the 10s NIST-FRVT requirement for a one-in-a-million search. In this paper, we present a fast and accurate biometric search solution under encryption for biometric data represented as vectors, a commonly used compact representation for compressing high-dimensional data. Our search solution is exhaustive and does not require any pre-selection (e.g., indexing) or dimensionality reduction techniques, which can degrade search accuracy. Also, it is parallelizable and supports dynamic databases. Our efficiency gain stems from our chunking and vertical organization of the encrypted reference database. This enables us to optimally leverage the single-instruction multiple-data (SIMD) property of FHE schemes. For a database of one million encrypted references represented by 512-dimensional vectors and a security level of 128 bits, our solution runs a fully encrypted exhaustive search in 9.23s measured on a 64-bit computer Intel Xeon Platinum 8358 with 64 CPUs and 64GB RAM. Hence, our solution is the first to satisfy the 10s NIST-FRVT limit under encryption, outperforming the state-of-the-art solutions by more than two orders of magnitude (at least 246 times faster) while achieving an accuracy improvement of at least 4.6% FNIR for a fixed FPIR at 0.1%.</p
Self-supervised learning of speech representations with Dutch archival data
Get PDFThis paper explores the use of Dutch archival television broadcast data for self-supervised learning of speech foundation models, specifically wav2vec 2.0. We first study data quality assumptions for pre-training, and show how music, noise and speaker overlap affect SSL convergence and downstream fine-tuning performance. Secondly, we explore effectively pre-processing strategies to convert the noisy broadcast dataset into a qualitative dataset for pre-training, by using Whisper and WhisperX. Thirdly, we compare mono-lingual and multi-lingual pre-training with equivalent amounts of data, and show that mono-lingual pre-training is more robust to out-of-domain data. Lastly, we achieve a state-of-the-art LARGE wav2vec 2.0 model for the Dutch language, by a continuation of pre-training a wav2vec 2.0 XLS-R model checkpoint with our 55k hour archival dataset
Detecting the invisible enemy in maize:Machine learning classification of fall armyworm damage in maize
Get PDFThe fall armyworm (FAW), Spodoptera frugiperda, is a highly destructive pest of maize, exploiting the crop’s nutritional and structural characteristics and causing significant yield loss. This study investigates the potential of remote sensing to detect and classify FAW infestations in Bangladesh maize fields using freely available Sentinel-1 (radar backscatter) and Sentinel-2 (optical reflectance) satellite datasets. Field observations were conducted during the 2019–2020 maize growing season in maize fields across six administrative divisions, covering both infested and non-infested sites across various maize growth stages. The corresponding Sentinel images were downloaded and processed for analysis. Radar backscatter values, spectral reflectance profiles, and vegetation indices were extracted from the Sentinel data, and statistical tests were performed to identify significant differences between infested and non-infested maize fields. Machine learning models were used to classify infestation severity based on five different combinations of Sentinel data and vegetation indices as inputs. Additionally, the results were validated through cross-validation techniques to ensure model accuracy. Statistical tests revealed significant differences between infested and non-infested maize across growth stages. Infested fields showed reduced near-infrared reflectance (Sentinel-2) and distinct radar backscatter patterns (Sentinel-1 VH polarisation), with notable variations at silking and maturity stages. The Random Forest classifier demonstrated superior accuracy and robustness, especially when integrating multi-source data. The red edge (740 nm) and near-infrared (865 nm) bands were particularly effective in distinguishing infestation levels, while the combination of spectral and radar data significantly improved classification accuracy by leveraging their complementary strengths. These results underscore the potential of integrating multi-source remote sensing data for scalable and accurate pest monitoring. Freely available Sentinel satellite imagery is a valuable source of information for early pest detection and management, aiding policymakers in identifying high-risk areas, implementing timely interventions, and promoting sustainable pest management strategies to protect maize production and reduce economic losses
Integrated Brillouin photonics in thin-film lithium niobate
Get PDFStimulated Brillouin scattering (SBS) is revolutionizing low-noise lasers and microwave photonics. However, a scalable and efficient integrated platform for Brillouin photonics has remained elusive. Here, we leverage the well-established thin-film lithium niobate (TFLN) platform to address these long-standing limitations. We report two distinct SBS processes on this platform, driven by surface acoustic wave (SAW) with 20-megahertz linewidth or bulk acoustic waves with a linewidth 200 times broader. Exploiting the strong SAW SBS gain, we demonstrate a narrowband internal net gain amplifier overcoming propagation losses. In addition, we achieve a stimulated Brillouin laser in TFLN, featuring a tuning range exceeding 20 nanometers and enabling high-purity radio frequency signal generation with an intrinsic linewidth of 9 hertz. Furthermore, we develop a programmable, multifunctional integrated Brillouin microwave photonic processor capable of notch filtering, bandpass filtering, or true time delay. This work bridges SBS with advanced TFLN technologies such as high-speed modulators and wide-band optical frequency combs, unlocking new paradigms for integrated Brillouin photonics.</p
Quantifying the Heat Load to Conductors due to Strain Energy Release in CTD-101K Magnet Impregnant
Get PDFState-of-the-art Nb 3Sn superconducting accelerator magnets are still prone to lengthy training. This study investigates whether direct heat from fractures in commonly used magnet impregnant is sufficient to cause the instabilities during training. To do so, the strain energy release rate of CTD-101 K epoxy is measured in liquid nitrogen at 77 K, together with the temperature of a nearby copper element representing a Nb 3Sn strand in the magnet windings. This experiment provides evidence that a large part of the mechanical energy is converted into heat. An analytical model using strain energy release rate data of impregnates at 4.2 K shows that this heat can contribute to the first quenches during magnet training.</p
Strategies for Antithrombotic Management During Non-cardiac Arterial Procedures:Results of the International ACTION Survey
Get PDFObjective: Peri-procedural antithrombotics are used extensively to prevent thromboembolic complications during non-cardiac arterial procedures (NCAP) worldwide. However, there is a lack of evidence to support recommendations on antithrombotic strategies, possibly leading to substantial variation in local practices. A comprehensive overview of antithrombotic strategies is needed to identify the most widely accepted protocols employed during NCAP, highlight variations in local practices, and identify new research targets to establish evidence based peri-procedural anticoagulation management. Methods: An international, web based survey study was conducted from March to October 2023, targeting vascular clinical specialists who applied antithrombotic strategies during NCAP in daily practice. Results: The survey was completed by 436 vascular clinical specialists from 45 countries (Europeans: 93%, vascular surgeons or vascular surgery residents: 98%). Systemic unfractionated heparin was used by nearly all vascular specialists during all procedures (varying between 98–99%, depending on the procedure type), but could vary depending on specific NCAP. A fixed starting dose (39–52%, most often 5 000 IU [80–89%]) or an actual bodyweight dependent dose (42–52%, most commonly 100 IU/kg [40–67%] or 50 IU/kg [17–40%]) was mainly used. Except during fenestrated or branched endovascular aneurysm repair procedures (51%), activated clotting time (ACT) was employed by a minority (26–31%). A large variety in measurement protocols was observed, yet a target ACT of 200 seconds was most often used for all NCAP types (44–54%). Most vascular specialists considered a heparin follow up dose (61–81%) and heparin reversal using protamine (54–63%), both for a variety of indications. Of the participants, 68% expressed discontent with their current antithrombotic protocol(s). Conclusion: This comprehensive, international survey study revealed large variation among vascular clinical specialists’ heparinisation strategies during NCAP. Together with the considerable discontent expressed regarding protocols, this emphasises the urgent need for comparative, randomised studies on antithrombotic management during NCAP.</p
Artifacts in photoacoustic imaging: Origins and mitigations
Get PDFPhotoacoustic imaging (PAI) is rapidly moving from the laboratory to the clinic, increasing the need to understand confounders which might adversely affect patient care. Over the past five years, landmark studies have shown the clinical utility of PAI, leading to regulatory approval of several devices. In this article, we describe the various causes of artifacts in PAI, providing schematic overviews and practical examples, simulated as well as experimental. This work serves two purposes: (1) educating clinical users to identify artifacts, understand their causes, and assess their impact, and (2) providing a reference of the limitations of current systems for those working to improve them. We explain how two aspects of PAI systems lead to artifacts: their inability to measure complete data sets, and embedded assumptions during reconstruction. We describe the physics underlying PAI, and propose a classification of the artifacts. The paper concludes by discussing possible advanced mitigation strategies
Age and orthostatic hypotension are associated with baroreflex sensitivity and cerebral oxygenation after postural change
No full textBaroreflex sensitivity (BRS), maintaining blood pressure (BP), and cerebral autoregulation, maintaining cerebral blood flow (CBF), are regulatory mechanisms to counteract posture-related BP changes and their effect on CBF. These mechanisms may fail in geriatric conditions such as orthostatic hypotension (OH) and cause symptoms and falls. This study aimed to determine the association of age, sex, antihypertensive use, comorbidity, and OH with BRS and cerebral oxygenation after postural change. Thirty-four younger adults (median age 25 years), 30 older adults (median age 77 years), and 41 geriatric outpatients (median age 76 years) performed 2–3 supine-stand transitions, while heart rate (electrocardiogram), BP (volume-clamp photoplethysmography), and cerebral oxygenation (near-infrared spectroscopy) were measured continuously. BRS, cerebral oxygenation and cerebral autoregulation were determined in the time and frequency domain. Associations were investigated using linear regression and group comparisons. Higher age and presence of OH (OH in at least one supine-stand transition during a continuous BP measurement) were associated with lower BRS (1 % per year, 30 % when having OH). Higher age was associated with higher cerebral oxygenation recovery (0.1 μmol/L per year) after 30 s, while OH was associated with lower cerebral oxygenation recovery (1.3 μmol/L when having OH) at 1 min after postural change. No evidence of cerebral autoregulation impairment was found across all three groups. Reduced BRS and cerebral oxygenation recovery specifically in participants with OH are in line with their assumed susceptibility to cerebral hypoxia. The role of cerebral autoregulation as a compensatory mechanism for failing BRS could not be confirmed.</p
Cost-effectiveness of [<sup>18</sup>F]FDG PET/CT in follow-up after thermal ablation in patients with colorectal liver metastases in the Dutch healthcare setting
Get PDFIntroduction: The ESMO and ECIO-ESOI consensus guidelines vary in their recommendations regarding the use of 18F-Fluorodeoxyglucose ([18F]FDG) PET/CT in the follow-up after thermal ablation in patients with colorectal liver metastases. This is partly because studies providing data on long-term benefits of [18F]FDG PET/CT are lacking. Therefore, a simulation model was developed to examine how follow-up with [18F]FDG PET/CT impacts treatment planning, health and cost outcomes. Methods: For an illustrative Dutch cohort, lifetime health and cost outcomes were simulated to assess the cost-effectiveness of performing a single additional [18F]FDG PET/CT. Patients followed a standard surveillance schedule consisting of three-monthly serum CEA and contrast-enhanced CT, plus [18F]FDG PET/CT 3–4 months after thermal ablation. Therapy could be repeated downstream the care pathway. Quality-of-life and survival estimates were based on disease stage and age. Costs were determined from a healthcare perspective incorporating costs related to diagnostics and treatments. The Consolidated Health Economic Evaluation Reporting Standards were followed. Results: Health benefits of additional [18F]FDG PET/CT were negligible, incremental QALYs < 0.001, whereas costs increased by €1,277, mainly due to the additional imaging. This lack of health benefits can be explained by the small subset of simulated patients (<5 %) in whom [18F]FDG PET/CT affected treatment planning. Discussion: Additional [18F]FDG PET/CT 3–4 months after thermal ablation is unlikely to be cost-effective. More research is needed to determine if using [18F]FDG PET/CT in subgroups of patients, or at alternative time points, is cost-effective. This requires collecting more (extensive) follow-up data across multiple centres to reflect heterogeneity between hospitals’ clinical practices.</p