Search

Climate change in the Mediterranean region is manifesting itself as an increase in average air temperature and a change in the rainfall regime: the value of cumulative annual rainfall generally appears to be constant, but the intensity of annual rainfall maxima, between 1 and 24 h, is increasing, especially in the period between late summer and early autumn. The associated ground effects in urban areas consist of flash floods and pluvial floods, often in very small areas, depending on the physical-geographical layout of the region. In the context of global warming, it is therefore important to have an adequate monitoring network for rain events that are highly concentrated in space and time. This research analyses the meteo-hydrological features of the 27 and 28 August 2023 event that occurred in the city of Genoa, Italy, just 4 d after the record maximum air temperature was recorded: between 19:00 and 02:00 UTC almost 400 mm of rainfall was recorded in the eastern sector of the historic centre of Genoa, with significant ground effects such as flooding episodes and the overflowing of pressurised culverts. Rainfall observations and estimates were made using both official or “authoritative” networks (rain gauges and meteorological radar) and rain gauge networks inspired by citizen science principles. The combined analysis of observations from authoritative and citizen science networks reveals, for the event analysed, a spatial variability of the precipitation field at an hourly and a sub-hourly timescale that cannot be captured by the current spatial density of the authoritative measurement stations (which have one of the highest densities in Italy). Monthly total rainfall and short-duration annual maximum time series recorded by the authoritative rain gauge network of the Genoa area are then analysed. The results show significant variation even at distances of less than 2 km in the average rainfall depth accumulated over sub-hourly duration. Extreme weather monitoring activity is confirmed as one of the most important aspects in terms of flood prevention and protection in urban areas. The integration between authoritative and citizen science networks can prove to be a valid contribution to the monitoring of extreme events.

Sentiment analysis (SA), also known as opinion mining, is a natural language processing (NLP) technique used to determine the sentiment or emotional tone behind a piece of text. It involves analyzing the text to identify whether it expresses a positive, negative, or neutral sentiment. SA can be applied to various types of text data such as social media posts, customer reviews, news articles, and more. This experiment is based on the Internet Movie Database (IMDB) dataset, which comprises movie reviews and the positive or negative labels related to them. Our research experiment’s objective is to identify the model with the best accuracy and the most generality. Text preprocessing is the first and most critical phase in an NLP system since it significantly impacts the overall accuracy of the classification algorithms. The experiment implements unsupervised sentiment classification algorithms including Valence Aware Dictionary and sentiment Reasoner (VADER) and TextBlob. We also examine the supervised sentiment classifications methods such as Naïve Bayes (Bernoulli NB and Multinomial NB). The Term Frequency-Inverse Document Frequency (TFIDF) model is used to feature selection and extractions. The combination of Multinomial NB and TFIDF achieves the highest accuracy, 87.63%, for both classification reports based on our experiment result.

Named-entity recognition (NER) is a crucial task in natural language processing, especially for extracting meaningful information from unstructured text data. In the healthcare domain, accurate NER can significantly enhance patient care by enabling efficient extraction and analysis of clinical information. This paper presents MedNER, a novel service-oriented framework designed specifically for medical NER in Chinese medical texts. MedNER leverages advanced deep learning techniques and domain-specific linguistic resources to achieve good performance in identifying diabetes-related entities such as symptoms, tests, and drugs. The framework integrates seamlessly with real-world healthcare systems, offering scalable and efficient solutions for processing large volumes of clinical data. This paper provides an in-depth discussion on the architecture and implementation of MedNER, featuring the concept of Deep Learning as a Service (DLaaS). A prototype has encapsulated BiLSTM-CRF and BERT-BiLSTM-CRF models into the core service, demonstrating its flexibility, usability, and effectiveness in addressing the unique challenges of Chinese medical text processing.

A procedure to adjust rainfall intensity (RI) measurements to account for the wind-induced measurement bias of traditional catching-type gauges is proposed and demonstrated with an application to a suitable case study. The objective is to demonstrate that adjustment curves derived from numerical simulations and disdrometer measurements allow for a posteriori correction of rainfall time series based on the wind velocity measurements alone. To quantify the impact of wind on long-term records, 1 min RI measurements from a cylindrical tipping-bucket rain gauge installed at the Hong Kong Observatory are adjusted. Catch ratios derived from the instrument's aerodynamic behaviour under varying wind speed and drop size combinations are obtained by fitting computational fluid-dynamic simulation results already available in the literature. Co-located high-resolution wind speed measurements from a cup and vane sensor and drop size distribution measurements from an optical video disdrometer (the 2DVD or two-dimensional video disdrometer) are used to infer the collection efficiency of the gauge as a function of wind speed and RI alone and to adjust raw data from a 4-year dataset (2018–2021) of 1 min RI measurements. Due to the specific local climatology, where strong wind is often associated with intense precipitation, 80 % of the dataset adjustments are limited to 4 % of the observed RI values. This, however, results in a significant amount of available freshwater resources that would be missing from the calculated hydrological and water management budget of the region should the adjustments be neglected. This work raises the need to quantify the impact of the wind-induced bias at other sites where disdrometer data support a characterisation of the relationship between the drop size distribution and the measured RI. Depending on the local rain and wind climatology, the correction may account for a significant portion of the annual rainfall amount.

Abstract Nonhealing wounds place a significant burden on both quality of life of affected patients and health systems. Skin substitutes are applied to promote the closure of nonhealing wounds, although their efficacy is limited by inadequate vascularization. The stromal vascular fraction (SVF) from the adipose tissue is a promising therapy to overcome this limitation. Despite a few successful clinical trials, its incorporation in the clinical routine has been hampered by their inconsistent results. All these studies concluded by warranting pre-clinical work aimed at both characterizing the cell types composing the SVF and shedding light on their mechanism of action. Here, we established a model of nonhealing wound, in which we applied the SVF in combination with a clinical-grade skin substitute. We purified the SVF cells from transgenic animals to trace their fate after transplantation and observed that it gave rise to a mature vascular network composed of arteries, capillaries, veins, as well as lymphatics, structurally and functionally connected with the host circulation. Then we moved to a human-in-mouse model and confirmed that SVF-derived endothelial cells formed hybrid human-mouse vessels, that were stabilized by perivascular cells. Mechanistically, SVF-derived endothelial cells engrafted and expanded, directly contributing to the formation of new vessels, while a population of fibro-adipogenic progenitors stimulated the expansion of the host vasculature in a paracrine manner. These data have important clinical implications, as they provide a steppingstone toward the reproducible and effective adoption of the SVF as a standard care for nonhealing wounds.

Coronavirus disease (COVID-19) is an infectious disease caused by SARS-CoV-2 virus, leading to mild to severe respiratory symptoms. Cardiovascular involvement is frequent and mainly manifests with myocarditis, arrhythmias, cardiac arrests, heart failure and coagulation abnormality. We generated human induced pluripotent stem cells (hiPSCs) from four COVID-19 patients, all characterized by increased levels of high-sensitivity Troponin I (hsTnI) during the infection acute phase, who developed (n = 2) or not (n = 2) severe myocarditis, as COVID-19 complication. The established hiPSCs were characterized for pluripotency and genomic stability, and constitute a useful resource for studying the mechanisms underlying the variability in COVID-19 severe cardiac manifestations.

Named-entity recognition (NER) is a crucial task in natural language processing, especially for extracting meaningful information from unstructured text data. In the healthcare domain, accurate NER can significantly enhance patient care by enabling efficient extraction and analysis of clinical information. This paper presents MedNER, a novel service-oriented framework designed specifically for medical NER in Chinese medical texts. MedNER leverages advanced deep learning techniques and domain-specific linguistic resources to achieve good performance in identifying diabetes-related entities such as symptoms, tests, and drugs. The framework integrates seamlessly with real-world healthcare systems, offering scalable and efficient solutions for processing large volumes of clinical data. This paper provides an in-depth discussion on the architecture and implementation of MedNER, featuring the concept of Deep Learning as a Service (DLaaS). A prototype has encapsulated BiLSTM-CRF and BERT-BiLSTM-CRF models into the core service, demonstrating its flexibility, usability, and effectiveness in addressing the unique challenges of Chinese medical text processing. [ABSTRACT FROM AUTHOR]

The increasing use of wireless communication and IoT devices has raised concerns about security, particularly with regard to attacks on the Routing Protocol for Low-Power and Lossy Networks (RPL), such as the wormhole attack. In this study, the authors have used the trust concept called PCC-RPL (Parental Change Control RPL) over communicating nodes on IoT networks which prevents unsolicited parent changes by utilizing the trust concept. The aim of this study is to make the RPL protocol more secure by using a Subjective Logic Framework-based trust model to detect and mitigate a wormhole attack. The study evaluates the trust-based designed framework known as SLF-RPL (Subjective Logical Framework-Routing Protocol for Low-Power and Lossy Networks) over various key parameters, i.e., low energy consumption, packet loss ratio and attack detection rate. The achieved results were conducted using a Contiki OS-based Cooja Network simulator with 30, 60, and 90 nodes with respect to a 1:10 malicious node ratio and compared with the existing PCC-RPL protocol. The results show that the proposed SLF-RPL framework demonstrates higher efficiency (0.0504 J to 0.0728 J out of 1 J) than PCC-RPL (0.065 J to 0.0963 J out of 1 J) in terms of energy consumption at the node level, a decreased packet loss ratio of 16% at the node level, and an increased attack detection rate at network level from 0.42 to 0.55 in comparison with PCC-RPL.

High-flow nasal cannula (HFNC) therapy is an oxygen delivery method particularly used in patients affected by hypoxemic respiratory failure. In comparison with the conventional “low flow” oxygen delivery systems, it showed several important clinical benefits. The possibility to nebulize drugs via HFNC represents a desirable medical practice because it allows the administration of inhaled drugs, mostly bronchodilators, without the interruption or modification of the concomitant oxygen therapy. HFNC, by itself has shown to exert a small but significant bronchodilator effect and improves muco-ciliary clearance; thus, the nebulization of bronchodilators through the HFNC circuit may potentially increase their pharmacological activity. Several technical issues have been observed which include the type of the nebulizer that should be used, its position within the HFNC circuit, and the optimal gas flow rates to ensure an efficient drug delivery to the lungs both in “quiet” and “distressed” breathing patterns. The aim of this review has been to summarize the scientific evidence coming from “in vitro” studies and to discuss the results of “in vivo” studies performed in adult subjects, mainly affected by obstructive lung diseases. Most studies seem to indicate the vibrating mesh nebulizer as the most efficient type of nebulizer and suggest to place it preferentially upstream from the humidifier chamber. In a quite breathing patterns, the inhaled dose seems to increase with lower flow rates while in a “distressed” breathing pattern, the aerosol delivery is higher when gas flow was set below the patient’s inspiratory flow, with a plateau effect seen when the gas flow reaches approximately 50% of the inspiratory flow. Although several studies have demonstrated that the percentage of the loaded dose nebulized via HFNC reaching the lungs is small, the bronchodilator effect of albuterol seems not to be impaired when compared to the conventional inhaled delivery methods. This is probably attributed to its pharmacological activity. Prospective and well-designed studies in different cohort of patients are needed to standardize and demonstrate the efficacy of the procedure.