Search

Cardiovascular diseases, including Heart Failure (HF), remain a leading global cause of mortality, often evading early detection. In this context, accessible and effective risk assessment is indispensable. Traditional approaches rely on resource-intensive diagnostic tests, typically administered after the onset of symptoms. The widespread availability of electrocardiogram (ECG) technology and the power of Machine Learning are emerging as viable alternatives within smart healthcare. In this paper, we propose several multi-modal approaches that combine 30-second ECG recordings and approximate long-term Heart Rate Variability (HRV) data to estimate the risk of HF hospitalization. We introduce two survival models: an XGBoost model with Accelerated Failure Time (AFT) incorporating comprehensive ECG features and a ResNet model that learns from the raw ECG. We extend these with our novel long-term HRVs extracted from the combination of ultra-short-term beat-to-beat measurements taken over the day. To capture their temporal dynamics, we propose a survival model comprising ResNet and Transformer architectures (TFM-ResNet). Our experiments demonstrate high model performance for HF risk assessment with a concordance index of 0.8537 compared to 14 survival models and competitive discrimination power on various external ECG datasets. After transferability tests with Apple Watch data, our approach implemented in the myHeartScore App offers cost-effective and highly accessible HF risk assessment, contributing to its prevention and management.

The detection and characterization of illegal solid waste disposal sites are essential for environmental protection, particularly for mitigating pollution and health hazards. Improperly managed landfills contaminate soil and groundwater via rainwater infiltration, posing threats to both animals and humans. Traditional landfill identification approaches, such as on-site inspections, are time-consuming and expensive. Remote sensing is a cost-effective solution for the identification and monitoring of solid waste disposal sites that enables broad coverage and repeated acquisitions over time. Earth Observation (EO) satellites, equipped with an array of sensors and imaging capabilities, have been providing high-resolution data for several decades. Researchers proposed specialized techniques that leverage remote sensing imagery to perform a range of tasks such as waste site detection, dumping site monitoring, and assessment of suitable locations for new landfills. This review aims to provide a detailed illustration of the most relevant proposals for the detection and monitoring of solid waste sites by describing and comparing the approaches, the implemented techniques, and the employed data. Furthermore, since the data sources are of the utmost importance for developing an effective solid waste detection model, a comprehensive overview of the satellites and publicly available data sets is presented. Finally, this paper identifies the open issues in the state-of-the-art and discusses the relevant research directions for reducing the costs and improving the effectiveness of novel solid waste detection methods.

This paper proposes a framework for developing forecasting models by streamlining the connections between core components of the developmental process. The proposed framework enables swift and robust integration of new datasets, experimentation on different algorithms, and selection of the best models. We start with the datasets of different issues and apply pre-processing steps to clean and engineer meaningful representations of time-series data. To identify robust training configurations, we introduce a novel mechanism of multiple cross-validation strategies. We apply different evaluation metrics to find the best-suited models for varying applications. One of the referent applications is our participation in the intelligent forecasting competition held by the United States Agency of International Development (USAID). Finally, we leverage the flexibility of the framework by applying different evaluation metrics to assess the performance of the models in inventory management settings.

We present a one-dimensional radiation-hydrodynamic model of a spherically symmetric cloud evolving under the influence of the self-gravity and the feedback from a star cluster forming in its centre. On one hand, the model is simple due to its 1D geometry, on the other hand, the feedback includes the ionising radiation, stellar winds and the radiation pressure acting on gas and dust. The star cluster is formed from the gas flowing into the cloud centre and the feedback parameters are determined from stellar evolution models and the cluster star forming history. The model is compared to the semi-analytic code WARPFIELD implementing similar physical processes and exploring the scenario that the young cluster R136 in the Large Magellanic Cloud was formed due to re-collapse of the shell formed by the previous generation star cluster. A good qualitative agreement is found, however, $3 - 4$ times higher stellar mass is needed to disrupt the cloud in our model, because it takes into account (contrary to WARPFIELD) self-gravity of the cloud surrounding the shell. We use the model to explore star formation in clouds with different mass, radius and density profile measuring their star formation efficiency (SFE), i.e. the fraction of the cloud mass converted to stars. We found that SFE is a function of a single parameter, $\mathrm{log(SFE)} \propto -n_{hm}^{-0.46}$, with $n_{hm}$ being the cloud mean particle density within its half-mass radius. Furthermore, we found that the feedback efficiency, i.e. a fraction of the feedback energy retained by gas, has a nearly constant value $\sim 10^{-3}$., Comment: 15 pages including appendix, 11 figures. Accepted for publication in MNRAS

In compact and dense star-forming clouds a global star cluster wind could be suppressed. In this case the stellar feedback is unable to expel the leftover gas from the cluster. Young massive stars remain embedded into a dense residual gas and stir it moving in the gravitational well of the system. Here we present a self-consistent model for the molecular gas distribution in such young, enshrouded stellar clusters. It is assumed that the cloud collapse terminates and the star formation ceases when a balance between the turbulent pressure and gravity and between the turbulent energy dissipation and regeneration rates is established. These conditions result in an equation that determines the residual gas density distribution that, in turn, allows one to determine the other characteristics of the leftover gas and the star formation efficiency. It is shown that model predictions are in good agreement with several observationally determined properties of cloud D1 in nearby dwarf spheroidal galaxy NGC 5253 and its embedded cluster., Comment: 12 pages, 3 figures, accepted for publication in the ApJ Letters

Survival modeling in healthcare relies on explainable statistical models; yet, their underlying assumptions are often simplistic and, thus, unrealistic. Machine learning models can estimate more complex relationships and lead to more accurate predictions, but are non-interpretable. This study shows it is possible to estimate hospitalization for congestive heart failure by a 30 seconds single-lead electrocardiogram signal. Using a machine learning approach not only results in greater predictive power but also provides clinically meaningful interpretations. We train an eXtreme Gradient Boosting accelerated failure time model and exploit SHapley Additive exPlanations values to explain the effect of each feature on predictions. Our model achieved a concordance index of 0.828 and an area under the curve of 0.853 at one year and 0.858 at two years on a held-out test set of 6,573 patients. These results show that a rapid test based on an electrocardiogram could be crucial in targeting and treating high-risk individuals., Comment: 4 pages, 4 figures

We present 3-D hydrodynamical models of the evolution of superbubbles powered by stellar winds and supernovae from young coeval massive star clusters within low metallicity ($Z = 0.02$Z$_{\odot}$), clumpy molecular clouds. We explore the initial stages of the superbubble evolution, including the occurrence of pair-instability and core-collapse supernovae. Our aim is to study the occurrence of dust grain growth within orbiting dusty clumps, and in the superbubble's swept-up supershell. We also aim to address the survival of dust grains produced by sequential supernovae. The model accounts for the star cluster gravitational potential and self-gravity of the parent cloud. It also considers radiative cooling (including that induced by dust) and a state-of-the-art population synthesis model for the coeval cluster. As shown before, a superbubble embedded into a clumpy medium becomes highly distorted, expanding mostly due to the hot gas streaming through low density channels. Our results indicate that in the case of massive ($\sim10^7$M$_{\odot}$) molecular clouds, hosting a super star cluster ($\sim5.6\times10^5$M$_{\odot}$), grain growth increments the dust mass at a rate $\sim4.8\times10^{-5}$M$_{\odot}$ yr$^{-1}$ during the first $2.5$Myr of the superbubble's evolution, while the net contribution of pair-instability and core-collapse supernovae to the superbubble's dust budget is $\sim1200$M$_{\odot} (M_{SC}/5.6\times10^{5}$M$_{\odot})$, where $M_{SC}$ is the stellar mass of the starburst. Therefore, dust grain growth and dust injection by supernovae lead to create, without invoking a top-heavy initial mass function, massive amounts of dust within low-metallicity star-forming molecular clouds, in accordance with the large dust mass present in galaxies soon after the onset of cosmic reionization., Comment: 8 pages, 4 figures. Accepted for publication in The Astrophysical Journal

Inspired by the excess soft X-ray emission recently detected in Green Pea galaxies, we model the soft X-ray emission (0.5 - 2.0 keV) of hot gas from star cluster winds. By combining individual star clusters, we estimate the soft X-ray emission expected from the typically unresolved diffuse hot gas in starburst galaxies, devoid of competing emission from e.g., AGN or other unresolved point sources. We use stellar models of sub-solar metallicities (0.02 $Z_{\odot}$ and 0.4 $Z_{\odot}$), and take into account supernova explosions for massive stars. For lower metallicities, we find that stellar winds do not contribute significantly ($\lesssim 3$ % of the mechanical energy) to the observed soft X-ray emission of normal star forming galaxies. For higher metallicities and possibly also for larger proportions of massive star clusters in the simulated starburst galaxies, we reproduce well the observed correlation between star formation rate and X-ray luminosity previously reported in the literature. However, we find that no combination of model assumptions is capable of reproducing the substantial soft X-ray emission observed from Green Pea galaxies, indicating that other emission mechanisms (i.e. unusually large quantities of High-/Low-Mass X-ray Binaries, Ultra-Luminous X-ray sources, a modified initial mass function, Intermediate-Mass Black Holes, or AGN) are more likely to be responsible for the X-ray excess., Comment: Accepted for publication in ApJ. 18 pages, 9 figures

We present a novel scenario for the growth of dust grains in galaxies at high-redshift ($z\sim 6$). In our model, the mechanical feedback from massive star clusters evolving within high-density pre-enriched media allows to pile-up a large amount of matter into massive supershells. If the gas metallicity ($\geq$ Z$_{\odot}$), number density ($\geq 10^6$ cm$^{-3}$) and dust-to-gas mass ratio ($\sim 1/150 \times Z$) within the supershell are sufficiently large, such supershells may become optically thick to the starlight emerging from their host star clusters and even to radiation from the Cosmic Microwave Background (CMB). Based on semi-analytic models, we argue that this mechanism, occurring in the case of massive ($\geq 10^7$ M$_{\odot}$) molecular clouds hosting $\geq 10^6$ M$_{\odot}$ star clusters, allows a large mass of gas and dust to acquire a temperature below that of the CMB, whereupon dust grain growth may occur with ease. In galaxies with total stellar mass $M_{*}$, grain growth within supershells may increase the dust mass by $\sim 10^6$ M$_{\odot}$ $(M_{*}/10^{8}$ M$_{\odot}$)., Comment: 10 pages, 7 figures, accepted for publication by MNRAS

Photon-assisted tunneling frequently provides detailed information on the underlying charge-transfer process. In particular, the Tien-Gordon approach and its extensions predict that the sideband spacing in bias voltage is a direct fingerprint of the number of electrons transferred in a single tunneling event. Here, we analyze photon-assisted tunneling into subgap states in superconductors in the limit of small temperatures and bias voltages where tunneling is dominated by resonant Andreev processes and does not conform to the predictions of simple Tien-Gordon theory. Our analysis is based on a systematic Keldysh calculation of the subgap conductance and provides a detailed analytical understanding of photon-assisted tunneling into subgap states, in excellent agreement with a recent experiment. We focus on tunneling from superconducting electrodes and into Yu-Shiba-Rusinov states associated with magnetic impurities or adatoms, but we also explicitly extend our results to include normal-metal electrodes or other types of subgap states in superconductors. In particular, we argue that photon-assisted Andreev reflections provide a high-accuracy method to measure small, but nonzero energies of subgap states which can be important for distinguishing conventional subgap states from Majorana bound states., Comment: 20 pages, 8 figures

Motivated by recent experiments searching for Majorana zero modes in tripartite semiconductor nanowires with epitaxial superconductor and ferromagnetic-insulator layers, we explore the emergence of topological superconductivity in such devices for paradigmatic arrangements of the three constituents. Accounting for the competition between magnetism and superconductivity, we treat superconductivity self consistently and describe the electronic properties, including the superconducting and ferromagnetic proximity effects, within a direct wave-function approach. We conclude that the most viable mechanism for topological superconductivity relies on a superconductor-semiconductor-ferromagnet arrangement of the constituents, in which spin splitting and superconductivity are independently induced in the semiconductor by proximity and superconductivity is only weakly affected by the ferromagnetic insulator., Comment: 10 pages, 6 figures

We discuss a theoretical model for the early evolution of massive star clusters and confront it with the ALMA, radio and infrared observations of the young stellar cluster highly obscured by the molecular cloud D1 in the nearby dwarf spheroidal galaxy NGC 5253. We show that a large turbulent pressure in the central zones of D1 cluster may cause individual wind-blown bubbles to reach pressure confinement before encountering their neighbors. In this case stellar winds are added to the hot shocked wind pockets of gas around individual massive stars that leads them to meet and produce a cluster wind in time-scales less than $10^5$ yrs. In order to inhibit the possibility of cloud dispersal, or the early negative star formation feedback, one should account for mass loading that may come, for example, from pre-main sequence (PMS) low-mass stars through photo-evaporation of their proto-stellar disks. Mass loading at a rate in excess of 8$\times 10^{-9}$ M$_{\odot}$ yr$^{-1}$ per each PMS star is required to extend the hidden star cluster phase in this particular cluster. In this regime, the parental cloud remains relatively unperturbed, while pockets of molecular, photoionized and hot gas coexist within the star forming region. Nevertheless, the most likely scenario for cloud D1 and its embedded cluster is that the hot shocked winds around individual massive stars should merge at an age of a few millions of years when the PMS star proto-stellar disks vanish and mass loading ceases that allows a cluster to form a global wind., Comment: 12 pages, 10 figures, accepted for publication in MNRAS

This paper proposes a new model based on Fuzzy k-Nearest Neighbors for classification with monotonic constraints, Monotonic Fuzzy k-NN (MonFkNN). Real-life data-sets often do not comply with monotonic constraints due to class noise. MonFkNN incorporates a new calculation of fuzzy memberships, which increases robustness against monotonic noise without the need for relabeling. Our proposal has been designed to be adaptable to the different needs of the problem being tackled. In several experimental studies, we show significant improvements in accuracy while matching the best degree of monotonicity obtained by comparable methods. We also show that MonFkNN empirically achieves improved performance compared with Monotonic k-NN in the presence of large amounts of class noise., Comment: Accepted in Neurocomputing