107036 research outputs found
Sort by
A meta-analysis towards the effectiveness of startup accelerators
No full textAccelerators have become crucial facilitators of entrepreneurial growth, yet empirical research on their effectiveness remains fragmented and inconclusive. To this extent, this paper conducts a comprehensive meta-analysis to synthesize quantitative findings and assess the impact of accelerator participation on new venture performance. By analyzing 21 primary studies and 68 effect sizes, the results reveal a statistically significant positive effect of accelerator participation. The study identifies key moderating factors, such as program duration, cohort size, sponsorship type, and regional context, which significantly influence venture outcomes. Additionally, the analysis highlights the presence selection biases within accelerator programs. This paper contributes to the theoretical understanding of accelerators by offering insights into their role in enhancing financial and operational dimensions of new venture performance. The findings provide valuable implications for both scholars and practitioners, laying the groundwork for future research on accelerator design and effectiveness
Robust and efficient pre-processing techniques for particle-based methods including dynamic boundary generation
No full textObtaining high-quality particle distributions for stable and accurate particle-based simulations poses significant challenges, especially for complex geometries. We introduce a preprocessing technique for 2D and 3D geometries, optimized for smoothed particle hydrodynamics (SPH) and other particle-based methods. Our pipeline begins with the generation of a resolution-adaptive point cloud near the geometry's surface employing a face-based neighborhood search. This point cloud forms the basis for a signed distance field, enabling efficient, localized computations near surface regions. To create an initial particle configuration, we apply a hierarchical winding number method for fast and accurate inside-outside segmentation. Particle positions are then relaxed using an SPH-inspired scheme, which also serves to pack boundary particles. This ensures full kernel support and promotes isotropic distributions while preserving the geometry interface. By leveraging the meshless nature of particle-based methods, our approach does not require connectivity information and is thus straightforward to integrate into existing particle-based frameworks. It is robust to imperfect input geometries and memory-efficient without compromising performance. Moreover, our experiments demonstrate that with increasingly higher resolution, the resulting particle distribution converges to the exact geometry
Empowering adolescents and young adults with somatoform disorders: a longitudinal pilot randomized controlled trial of a parent-focused mindfulness-based training versus support groups
No full textBackground
In childhood somatoform disorders develop and persist through bio-psycho-social processes including family variables, substantially impacting the development of children, their school attendance, and well-being. Although intensive interdisciplinary pain treatment (IIPT) has demonstrated effectiveness, maintaining long-term outcomes remains challenging, particularly given children's increased risk of developing affective disorders. Since family dynamics influence disorder development and persistence, this study aimed to investigate the potential added value of addressing parents within the therapeutic process.
Method
This monocentric longitudinal pilot randomized controlled trial (RCT) (N = 31) investigated adjunct parent-focused interventions (mindfulness vs. support group) for families of children with somatoform disorders enrolled in IIPT. We assessed children's empowerment, pain related disability, affective pain perception, pain characteristics and parental catastrophizing at admission, discharge, and six-month follow-up. Regression models and repeated-measures ANOVAs evaluated intervention effects.
Results
Multiple regression analysis identified older age as predictor of empowerment at baseline (β = 0.63, p < .01). IIPT plus parental intervention showed lasting treatment effects. Group allocation had no effect on empowerment, pain disability, affective pain perception, or parental catastrophizing, but allocation to the support group (B = -2.49, p = .009) and male sex (B = -2.29, p = .008) predicted greater reduction in average pain intensity.
Conclusions
This pilot RCT showed no superiority of the parent-focused mindfulness group over the support group when combined with IIPT. Yet, both interventions may enhance treatment efficacy, with some evidence suggesting additional benefits of the support group intervention in reducing average pain intensity. Further research is needed to examine underlying mechanisms and establish optimal parent intervention approaches in paediatric pain treatment
Die Entwicklung von Künstliche Intelligenz-Systemen für die Diagnostik und Therapie von Frühneoplasien im oberen Gastrointestinaltrakt am Beispiel des Barrett-Karzinoms: Grenzen und Möglichkeiten
No full textIm Rahmen dieser Arbeit erfolgte die Entwicklung und Testung eines KI-Systems, welches eine pixelgenaue Segmentierung von BE bzw. EAC in Echtzeit erlaubt. In der durchgeführten randomisierten Tandem-Studie an standardisierten Videos von BE war die eigenständige Performance des KI-Systems vergleichbar gut wie die der ebenfalls die Studie absolvierenden BE-Expert*innen.
Darüber hinaus zeigte der Einsatz von KI einen deutlichen Zusatznutzen in der Gruppe der nicht-BE-Expert*innen. Nicht-BE-Expert*innen mit der Unterstützung von KI zeigten eine signifikant bessere Performance als ohne. Der Einsatz von KI als Entscheidungsunterstützungssystem führte zu einer höheren Konfidenz der Endoskopiker*innen hinsichtlich der Richtigkeit der getroffenen Entscheidung. Eine höhere diagnostische Konfidenz war mit einer verbesserten Performance assoziiert.
Als nächsten Schritt sind klinische Studien zur Untersuchung des Effekts von KI auf die Performance von Endoskopiker*innen bei der Evaluation von BE wichtig. Des Weiteren ist eine weitere Erforschung der Mensch-Maschine-Interaktion zur optimalen Implementierung von KI-Anwendungen in den klinischen Alltag essenziell.In conclusion in this study an AI-system, that allows real-time segmentation of BE and EAC, was developed and tested. We conducted a randomized Tandem Trial using standardized endoscopy videos of BE. In this trial the standalone performance of the AI-system was comparable to the performance of expert endoscopists. Furthermore, especially non-BE-experts benefitted from additional AI as clinical decision support system. Non-BE-experts using AI as clinical decision support system improved their performance significantly compared to when not using it. Application of AI as clinical decision support system improved diagnostic confidence of endoscopists. Higher diagnostic on the other hand was associated with improved performance during the Evaluation of BE. Additional studies are necessary to evaluate the effect of AI as decision support system on the performance of endoscopists in clinical practice. Furthermore, ideal implementation of AI into clinical practice requires more studies to better understand human-computer interaction while using it as clinical decision support system
Updating the global weighting factor for biodiversity impact assessment in LCA
No full textLife‑cycle impact assessment (LCIA) is progressively expanding to address biodiversity impacts, but spatial heterogeneity continues to dominate the associated uncertainty. To account for this, the Ecoregion Factor (EF)—a dimensionless weighting coefficient—modulates biodiversity‑impact scores according to the ecological richness of the ecoregion in which a land‑use takes place. Since the initial publication, the underlying spatial datasets have been superseded by more recent releases. Consequently, we recomputed the EF using the latest ecoregional delineation and the most recent global layers for grasslands, forests and wetlands. To facilitate integration with LCIA frameworks that operate at the national level, the updated EF values were aggregated to country‑level averages by calculating an area‑weighted mean of the normalised EF across all ecoregions intersecting each nation’s borders. This yields a robust geographic adjustment factor that preserves the spatial nuance of ecoregional biodiversity while making biodiversity impact assessments feasible even when inventory data are available only at the country scale. By providing an up‑to‑date, geographically calibrated weighting coefficient, the revised EF enhances the spatial granularity of biodiversity‑focused LCIA results
Chiral and topological spin textures and their ultrafast dynamics in thin-film materials with local inhomogeneities
No full textMagnetic thin films with perpendicular anisotropy are widely studied as they exhibit a broad spectrum of magnetic domain textures and compelling properties for future spintronic and neuromorphic computing applications. A key factor influencing domain behavior is the chirality of the surrounding domain walls, which governs their stability, mobility, and topology.Since early research into bubble domain materials in the 1960s, the properties and dynamics of domain walls have been studied continuously.Although it has long been known that pinning at material defects affects domain walls, the ability to study these interactions at the microscopic level has been limited.Steady technological advancement in ultrafast excitation techniques and x-ray imaging now enables new insights into these nanometer-scale magnetic textures.
In this thesis, we examine how material defects and inhomogeneities affect the statics and dynamics of chiral and topological spin structures. We focus on three case studies, each targeting a distinct aspect of defect-induced magnetic behavior. For each project, we develop and deploy x-ray imaging and characterization techniques based on circular and linear magnetic x-ray dichroism contrast to resolve features spatially on the nanometer scale and with picosecond time resolution, supported by complementary measurements and modeling.
First, we investigate the impact of lateral inhomogeneities on the static domain wall chirality in DyCo thin films. Our vector imaging technique reveals strong lateral chirality variations imprinted by the local material properties, which appear to be influenced by the initial as-grown domain state.
Second, we utilize vector x-ray imaging to analyze domain wall defects in all-optically switched GdFe samples.
Analysis of the defect density allows insights into the switching and nucleation dynamics of the domain wall itself.
We uncover that post-nucleation domain wall propagation through an inhomogeneous material matrix is the dominant nucleation channel for domain wall defects.
Third, we study the laser-induced nucleation and localization of magnetic skyrmions in a patterned Co/Pt multilayer. We consolidate the homogeneous, exchange-driven nucleation of skyrmions after laser excitation with the observation that local anisotropy variations can induce a deterministic localization: We find that the localization is determined by purely local stability characteristics that induce skyrmion decay or proliferation.
Our findings collectively demonstrate how lateral material inhomogeneities govern both equilibrium configurations and dynamic processes of chiral and topological magnetic textures, offering key insights for future device design and fundamental studies of magnetic matter
Mobile interfaces for caregivers and older adults: iterative design of the LifeTomorrow Ecosystem with aesthetic and functional considerations
No full textAs the population of older adults increases, so does the demand for technology that supports caregiving and aging in place. Smart home technology, wearable health trackers, and mobile applications have all been identified as possible methods of support. Studies on the user interfaces of these technologies have predominantly explored how well their features and functions address the complex needs of older adults and caregivers. However, many of these applications lack adequate consideration of visual design principles and aesthetics. The present study aims to illustrate the iterative design process of the LifeTomorrow Ecosystem which includes two applications: one for caregivers and one for older adult care recipients. The results include high-fidelity screens from the applications that incorporate functional and visual design principles, as well as the feedback of older adults, caregivers, and designers. Finally, we provide recommendations for designers to consider when designing applications targeted at older adults and their caregivers
