275003 research outputs found
Sort by
Biodegradable nanoparticles for specific drug transport
No full text255274Biodegradable nanoparticles are highly versatile and adaptable delivery systems for pharmaceutical drugs. The nanoparticles can be targeted to specific cell types and tissues, modified to overcome biological barriers the contained drugs cannot overcome on their own and at the same time mask the undesirable side effects of pharmacological substances. The nanoparticle characteristics can be tailored to specific targets by modifications such as the addition of antibodies to the particle surface or coating them in substances altering their circulation behaviour. Different approaches and examples for targeted drug delivery with biodegradable nanoparticles are discussed, as well as model systems for the advanced evaluation of the used formulations shown
Recyclingfähigkeit von Permanentmagnet-Synchronmotoren zur Reduzierung der Materialkritikalität
No full text5054851
PSIPRO - Praxistransfer skalierbarer innovativer Produkt- und Prozesslösungen
No full textDer Bericht enthält grundlegende Informationen zum Projek PSIPRO welches darauf abzielt, den Transfer innovativer Produkt- und Prozesslösungen im Bauwesen zu verbessern. Hinzu kommen Handlungsempfehlungen und Steckbriefe zu den Fokusthemen "Baustoffe (Beton und Stahl)", "Modulares Bauen", "Automatisierte Baustelle und Robotik" sowie "Serielle Sanierung"
Interpretable Topic Extraction and Word Embedding Learning Using Non-Negative Tensor DEDICOM
No full text123167Unsupervised topic extraction is a vital step in automatically extracting concise contentual information from large text corpora. Existing topic extraction methods lack the capability of linking relations between these topics which would further help text understanding. Therefore we propose utilizing the Decomposition into Directional Components (DEDICOM) algorithm which provides a uniquely interpretable matrix factorization for symmetric and asymmetric square matrices and tensors. We constrain DEDICOM to row-stochasticity and non-negativity in order to factorize pointwise mutual information matrices and tensors of text corpora. We identify latent topic clusters and their relations within the vocabulary and simultaneously learn interpretable word embeddings. Further, we introduce multiple methods based on alternating gradient descent to efficiently train constrained DEDICOM algorithms. We evaluate the qualitative topic modeling and word embedding performance of our proposed methods on several datasets, including a novel New York Times news dataset, and demonstrate how the DEDICOM algorithm provides deeper text analysis than competing matrix factorization approaches.3
A double-blind, placebo-controlled trial of the efficacy and safety of two doses of azelastine hydrochloride in perennial allergic rhinitis
No full textBackground: Azelastine hydrochloride (AZE) is a selective, non-sedating H1 antagonist with anti-inflammatory and mast cell stabilizing properties, which can be used as an alternative to intranasal corticosteroids. The objective of this study was to evaluate the efficacy of the new formulation of 0.15% AZE compared to that of the placebo at a dosage of two sprays per nostril twice daily for 4 weeks in patients with perennial allergic rhinitis (PAR). Materials and methods: A total of 581 subjects were randomized in this double-blind (DB) placebo-controlled trial (NCT00712920) that compared 0.10% (1,096 μg daily) and 0.15% AZE (1,644 μg daily) to the placebo in PAR patients. The study consisted of a 7-day single-blind placebo lead-in period and a 28-day DB treatment period. The primary endpoint was the change from baseline in the 12-h reflective total nasal symptom score (rTNSS) for the entire 28-day study period of 0.15% AZE, two sprays per nostril BID compared to the placebo. The efficacy and safety of 0.15% AZE were compared to the placebo. Results: Least square (LS) mean improvement from baseline in the morning (AM) and evening (PM) combined rTNSS was statistically significant for the 0.15% AZE group (p = 0.04) compared to the placebo group. LS mean improvement from baseline in the AM and PM combined rTNSS was 4.10 (4.26) units for 0.15% AZE and 3.81 (3.99) for 0.10% AZE. For individual symptoms, there was a statistically significant change in the LS mean (p = 0.04) improvement from baseline on the 12-h reflective assessment for the 0.15% AZE group for runny nose. Further numerical improvements were shown for itchy nose, nasal congestion, runny nose, and sneezing compared to the placebo. No deaths or serious adverse events related to the study medication were reported. Conclusion: The present formulation of 0.15% AZE is safe and effective in relieving PAR symptoms. It effectively relieves nasal and non-nasal symptoms. Clinical Trial Registration: ClinicalTrials.gov, identifier: NCT00712920.
Exploiting acquisition measurements and spatial processing for improved GNSS spoofing detection and classification with snapshot receivers
No full text34653495Spoofing attacks mislead global navigation satellite system (GNSS) receivers into reporting false position, velocity, and time (PVT) solutions. However, not all spoofing attacks are the same: they range from simple meaconers to advanced synchronized spoofers. Classifying the attack improves situational awareness and helps the receiver operator take appropriate counteractions. This paper presents spoofing detection and spoofing attack classification using lightweight spatial and acquisition metrics from a snapshot receiver. Further, appropriate machine learning (ML) techniques provide high performance and intuition. The results show that spatial information (i.e., multi-antenna comparisons) has superior spoofing detection, but incorporating the acquisition metrics (i.e., temporal information) facilitates spoofing attack type classification. Finally, only basic acquisition metrics are required for classification, demonstrating lightweight feature selection and training. This work shows that spoofing detection and attack type classification is possible and emphasizes that a low feature space is required
Institutional barriers to dynamic truck charging: why electric road systems struggle in Europe
No full textElectric road systems (ERS) have been proposed as an efficient solution to dynamically charge electric trucks but have not yet become a dominant solution. This paper provides an institutional explanation for the case of Europe, building on 22 expert interviews in eight European countries, event observations, and policy documents. The analysis identifies three main explanations. Firstly, as a line infrastructure, ERS require government commitment for build-up and coordination, particularly across borders. This conflicts with the widespread idea of technology-open governments that provide R&D funds to initiate market-driven solutions. Secondly, time constraints favour readily available solutions backed by industry, like stationary charging, over sector-specific ERS technologies that lack a unified lobby with policy access. Thirdly, ERS technologies challenge long-standing sectoral designs. If policymakers want to maintain the option of ERS alongside stationary charging, they need to acknowledge this institutional uphill battle and consider compatibility requirements for vehicles and an active commitment for larger routes.15
Image-to-Image Translation for Simplified MRI Muscle Segmentation
No full textDeep neural networks recently showed high performance and gained popularity in the field of radiology. However, the fact that large amounts of labeled data are required for training these architectures inhibits practical applications. We take advantage of an unpaired image-to-image translation approach in combination with a novel domain specific loss formulation to create an “easier-to-segment” intermediate image representation without requiring any label data. The requirement here is that the task can be translated from a hard to a related but simplified task for which unlabeled data are available. In the experimental evaluation, we investigate fully automated approaches for segmentation of pathological muscle tissue in T1-weighted magnetic resonance (MR) images of human thighs. The results show clearly improved performance in case of supervised segmentation techniques. Even more impressively, we obtain similar results with a basic completely unsupervised segmentation approach.
Lidar Emitter and Multi-species greenhouse gases Observation iNstrument (LEMON): advances on a multi-species differential absorption Lidar system
No full textIn the frame of LEMON project (Lidar Emitter and Multi-species greenhouse gases Observation iNstrument - European Union's Horizon 2020 research and innovation program, GA n°821868), we are developing a multi-species differential absorption Lidar (DIAL). The goal is to benefit from innovative technological developments in terms of optical emitter, spectral reference, to be able to address H2O and its isotope HDO at 1982 nm, CO2 at 2051 nm, and potentially CH4 at 2290 nm, for future ground-based range-resolved DIAL sensing, and with the prospect of future airborne integrated-path DIAL (IPDA). The infrared emitter is based on the combination of two specific, patented, no-seeder Nested Cavity OPOs (NesCOPOs) coupled to a single optical parametric amplifier (OPA) line for high energy pulses generation. Specific developments are also pursued on the frequency reference for the emitter, which is planned to be provided by a GPS referenced frequency comb against which the emitter output pulses can be heterodyned. Besides the instrument design, specific tests experiments have been carried out, covering a wide panel of activities: radiation testing of some critical components to assess the potential of some key components for future space applications, emitter and frequency reference testing, preliminary DIAL tests with laboratory test-beds and comparison with specific in-situ calibration instruments as well as additional innovative techniques evaluation for the emitter. The final instrument design was carried out and the sub-units are now being built
First demonstration of in-memory computing crossbar using multi-level Cell FeFET
No full textAdvancements in AI led to the emergence of in-memory-computing architectures as a promising solution for the associated computing and memory challenges. This study introduces a novel in-memory-computing (IMC) crossbar macro utilizing a multi-level ferroelectric field-effect transistor (FeFET) cell for multi-bit multiply and accumulate (MAC) operations. The proposed 1FeFET-1R cell design stores multi-bit information while minimizing device variability effects on accuracy. Experimental validation was performed using 28 nm HKMG technology-based FeFET devices. Unlike traditional resistive memory-based analog computing, our approach leverages the electrical characteristics of stored data within the memory cell to derive MAC operation results encoded in activation time and accumulated current. Remarkably, our design achieves 96.6% accuracy for handwriting recognition and 91.5% accuracy for image classification without extra training. Furthermore, it demonstrates exceptional performance, achieving 885.4 TOPS/W–nearly double that of existing designs. This study represents the first successful implementation of an in-memory macro using a multi-state FeFET cell for complete MAC operations, preserving crossbar density without additional structural overhead.14