14043 research outputs found
Sort by
Exploring smartphone-based edge AI inferences using real testbeds
No full textThe increasing availability of lightweight pre-trained models and AI execution frameworks is causing edge AI to become ubiquitous. Particularly, deep learning (DL) models are being used in computer vision (CV) for performing object recognition and image classification tasks in various application domains requiring prompt inferences. Regarding edge AI task execution platforms, some approaches show a strong dependency on cloud resources to complement the computing power offered by local nodes. Other approaches distribute workload horizontally, i.e., by harnessing the power of nearby edge nodes. Many of these efforts experiment with real settings comprising SBC (Single-Board Computer)-like edge nodes only, but few of these consider nomadic hardware such as smartphones. Given the huge popularity of smartphones worldwide and the unlimited scenarios where smartphone clusters could be exploited for providing computing power, this paper sheds some light in answering the following question: Is smartphone-based edge AI a competitive approach for real-time CV inferences? To empirically answer this, we use three pre-trained DL models and eight heterogeneous edge nodes including five low/mid-end smartphones and three SBCs, and compare the performance achieved using workloads from three image stream processing scenarios. Experiments were run with the help of a toolset designed for reproducing battery-driven edge computing tests. We compared latency and energy efficiency achieved by using either several smartphone clusters testbeds or SBCs only. Additionally, for battery-driven settings, we include metrics to measure how workload execution impacts smartphone battery levels. As per the computing capability shown in our experiments, we conclude that edge AI based on smartphone clusters can help in providing valuable resources to contribute to the expansion of edge AI in application scenarios requiring real-time performance
Optimizing immunization strategies for individuals living with HIV
No full textHuman immunodeficiency virus (HIV) infection remains a major challenge in global health. In recent years, vaccines have emerged as an important tool for the treatment and prevention of HIV-related complications. This review article addresses the evolving landscape of vaccines for people living with HIV (PLWH), evaluating current vaccination strategies for standard vaccines and travel vaccines in PLWH compared to the general population and offering a summary of the current recommended vaccines. It evaluates studies for vaccine effectiveness and safety and discusses methods to improve vaccination rates among PLWH. Systematic research was carried out using keywords. We address the current state of knowledge and highlight areas for future research and development
Theory of potential impurity scattering in pressurized superconducting
No full textRecently discovered high-T superconductivity in pressurized bilayer nickelate (\La_{3}Ni_{2}O_{7}\)(La-327)is likely driven by the non-phononic repulsive interaction. Depending on the interlayer repulsion strength, the superconducting gap structure is expected to be either -wave or sign-changing bonding-antibonding s-wave. Unfortunately, conventional spectroscopic probes of the gap structure are impractical due to the high-pressure requirement. We propose studying the effect of point-like non-magnetic impurities to distinguish these symmetries, which can be achieved by electron irradiation before applying pressure. Here, we theoretically predict conventional suppression for -wave superconductivity, whereas the suppression for the interlayer s -wave state depends subtly on the asymmetry of bonding and antibonding subspaces. For the predicted electronic structure of La-327, the s -wave is more robust, with T showing a convex-to-concave transition, indicating a crossover to s-wave symmetry as impurity concentration increases. We further analyze the sensitivity of these findings to potential electronic structure modifications
Alpha-synuclein misfolding as fluid biomarker for Parkinson’s disease measured with the iRS platform
No full textMisfolding and aggregation of alpha-synuclein (αSyn) play a key role in the pathophysiology of Parkinson’s disease (PD). Despite considerable advances in diagnostics, an early and differential diagnosis of PD still represents a major challenge. We innovated the immuno-infrared sensor (iRS) platform for measuring Syn misfolding. We analyzed cerebrospinal fluid (CSF) from two cohorts comprising PD cases, atypical Parkinsonian disorders, and disease controls. We obtained an AUC of 0.90 ( = 134, 95% CI 0.85–0.96) for separating PD/MSA from controls by determination of the Syn misfolding by iRS. Using two thresholds divided individuals as unaffected/affected by misfolding with an intermediate area in between. Comparing the affected/unaffected cases, controls versus PD/MSA cases were classified with 97% sensitivity and 92% specificity. The spectral data revealed misfolding from an -helical/random-coil Syn in controls to -sheet enriched Syn in PD and MSA cases. Moreover, a first subgroup analysis implied the potential for patient stratification in clinically overlapping cases. The iRS, directly measuring all Syn conformers, is complementary to the Syn seed-amplification assays (SAAs), which however only amplify seeding competent conformers
Body mass index is not associated with lymph node metastasis, infection, and seroma after sentinel lymph node biopsy in cutaneous melanoma
No full textImplementation of a digital nurse to improve the use of digital health applications (DiGA) for older people with depressive disorders (DiGA4Aged)
No full text
In the face of extensive waiting times for outpatient psychotherapy, prescriptible digital health applications (DiGA) are a useful and effective addition to the range of available therapy options for patients with mild to moderate depression. However, older adults face a particular challenge in implementing DiGA since higher age is a decisive predictor of lower digital health literacy. The necessity of an independent use of the prescribed DiGA is therefore associated with challenges for older patients and providers. In practice, it is crucial not to leave patients, especially older adults, alone after prescribing, but to maintain close contact to overcome technical and motivational barriers and to ensure that the novel application is used. However, this is difficult for physicians and psychotherapists due to the critical healthcare system situation in Germany described above. Another support system is needed. Hence, the main hypothesis of this study is that the additional implementation of digital nurses leads to a higher percentage of older patients with depressive symptoms starting DiGA use compared to a prescription and information alone.
Two DiGA for mild to moderate depression in older patients were available and permanently approved at the time of the funding application. Using the most suitable one of them, as shown in a pilot study, the feasibility of implementation will be examined within a randomized proof of concept study. In our study, a digital nurse is trained to support patients with depression in using a DiGA. The main outcome is DiGA use (first session started: yes/no) after 8 weeks. Major secondary outcomes are patient-relevant outcomes, feasibility of recruitment and intervention, and factors moderating the effect or predicting DiGA use in the target group. Best practice guidelines will be elaborated on how to support and improve DiGA prescription and successful use in this population.
In Germany, the approved DiGA are currently little used, especially by people with a low digital affinity. This proof of concept study will use the example of older people with depressive disorders to show whether it is possible to increase the usage rate of a DiGA with the support of a digital nurse so that a DiGA can become a serious therapy option
Impact of experimental inflammation on the neuronal processing of cardiac interoceptive signals and heart rate variability in humans
No full textInteroception, or the perception of internal somatic states, is crucial for signaling the individual to take care of the body when needed. It enables behavioral adaptations to sickness states, which further impact autonomic nervous system (ANS) activity. Whether acute inflammation affects interoceptive processing and how this relates to sickness behavior remains unknown. Therefore, we investigated interoceptive processing in participants undergoing experimental endotoxemia. In neuroimaging research, heartbeat-evoked potentials (HEP) - defined as event-related potentials time-locked to electrocardiogram (ECG) R-waves during electroencephalogram (EEG) recordings - have emerged as a promising metric for cardiac interoceptive processing. We analyzed the effects of intravenous administration of lipopolysaccharide (LPS; 0.4 ng/kg) or placebo, on HEP amplitudes and ANS functioning in healthy, female participants ( = 52) during 8 min resting-state EEG and ECG recordings before and 2 h after injections. Our results showed increased cortisol and cytokine levels in the LPS group, along with increased sympathetic and decreased parasympathetic activity 2 h after injections compared to the placebo group. Placebo-injected participants exhibited lower post injection-baseline differences in HEP amplitudes in an early timeframe (255–455 ms), indicating lower HEPs 2 h after administrations. Moreover, post-injection HEP amplitudes differed between groups, suggesting that while participants in the placebo group showed altered HEP amplitudes after injection, HEPs remained unresponsive to LPS administration. These findings are discussed in the context of predictive processing, expectation violation and attention direction to external and interoceptive cues. Future research should further investigate the role of LPS dose and explore behavioral measures of interoception under experimental inflammation
Anti-CD19 CAR T-cell therapy in advanced stiff-person syndrome and concomitant myasthenia gravis
No full text
Autologous chimeric antigen receptor (CAR) T-cell therapy has recently gained interest in the treatment of rheumatic and neuroimmunologic diseases.
We report a 62-year-old female patient with a 14-year history of treatment-refractory anti–GAD-positive stiff-person syndrome (SPS) and concomitant anti–AChR-positive myasthenia gravis. Despite a relatively stable disease course in the first 8 years, SPS dramatically progressed afterward. In 2023, she was able to walk less than 10–15 m and suffered from severe persistent stiffness in the left arm superimposed with painful muscle spasm attacks (MSAs). Numerous immunotherapies, including intravenous immunoglobulins, plasma exchange, steroids, azathioprine, and rituximab, were ineffective. Consequently, she was escalated to compassionate use of autologous anti-CD19 CAR T-cell therapy (KYV-101).
From the third month post–CAR-T, we observed a substantial improvement in walking distance, pain, anxiety, and MSAs in the arm. By the sixth month, she was able to walk 500 m. The anti-GAD titers declined from 1:320 to 1:32. Side effects included grade 2 cytokine release syndrome and moderate leukopenia, without serious infections.
CAR T-cell therapy was effective at mitigating SPS symptoms, despite the long history and severe refractory disease course in our patient. Controlled trials are needed to evaluate its potential in SPS
Formation of hydrogen trioxide (HOOOH) in extraterrestrial ice analogs and its role as an oxidizer in prebiotic chemistry
No full textThe formation and characterization of hydrogen trioxide have fascinated scientists for more than a century, due to its role as a prototype model for oxygen-chain bonding and as a key transient in antibody-catalyzed oxidation reactions relevant to the origin of life. However, the abiotic formation pathways to hydrogen trioxide have remained elusive. Here, we demonstrate in laboratory simulation experiments that hydrogen trioxide effectively forms in water–molecular oxygen ice analogs at temperatures as low as 5 kelvin under exposure to proxies of galactic cosmic rays. Exploiting synchrotron vacuum ultraviolet photoionization reflectron time-of-flight mass spectrometry, hydrogen trioxide along with hydrogen peroxide and the hydroperoxyl radical was identified during the temperature-programmed desorption of the irradiated ices. This abiotic synthesis expands the oxidant inventory on interstellar grains, icy moons, and Kuiper belt objects, offering a plausible source of essential oxidizers for prebiotic chemistry in space
Analysis and improvements of post-quantum cryptosystems
No full textPublic-Key Kryptographie ist das Fundament der modernen Kryptographie. Quantenalgorithmen können die gängigen klassischen Verfahren basierend auf dem RSA oder dem Diffie-Hellman Problem theoretisch brechen.
In dieser Arbeit werden bekannte, potentiell quantenresistente, Public-Key Verschlüsselungsverfahren verbessert und im Quantom Random Oracle Model (QROM) analysiert.
Das QROM ist die natürliche Erweiterung des Random Oracle Models wenn man die Resistenz gegen Quantenangreifer betrachtet.
Konkret behandelt diese Arbeit die aktive Multi-User Sicherheit von Verfahren basierend auf der Fujisaki-Okamoto Transformation.
Weiterhin werden in der Arbeit effizientere Varianten des Gitter-basierten Verfahrens NTRU vorgestellt und analysiert.
Zuletzt analysiert die Arbeit welche Sicherheitsannahmen notwendig und hinreichend sind um die Sicherheit von Nicht-Interaktiven Schlüsselaustauschverfahren basierend auf Gruppenwirkungen zu beweisen