A detailed study is presented in order to investigate the dynamics of compound vesicles confined in a channel under shear flow. Compound vesicles, which are made of a smaller vesicle embedded within a larger one, are systems of relevant biological importance. Indeed, they can be seen as biomimetic models for multi-compartmentalized cells such as leukocytes, nucleate cells, or vesicles embedding even more vesicles, as in the case of vesosomes, with several possible applications. Despite the relevance, in experiments it is extremely difficult to set independently all the parameters controlling the system, while available numerical studies neglect thermal fluctuations which are relevant in vesicle dynamics. The present study [1] adopts mesoscale hydrodynamic simulations of a model system, explicitly including thermal fluctuations. This allows the observations of a rich phenomenology in the dynamical behavior of compound vesicles, which well matches with the few available experiments [2,3]. Moreover, our study allows a full characterization of the undulating motion under different conditions to an extent never observed before. This latter dynamical state is characterized by periodic oscillation of the inclination and buckling of the external membrane, as illustrated in the figure.

Noise is a realistic source of corruption in dermoscopic imaging, stemming from sensor limitations, low-light acquisition, and uncontrolled capture conditions. This study examines how this type of degradation alters the behavior of CNN-based skin lesion classifiers by systematically perturbing a balanced ISIC 2018 subset with increasing noise intensities. Four representative CNN architectures are compared under a unified training protocol, and their performance is analyzed across evaluation setups that separate matched and mismatched noise conditions between training and inference. Beyond standard Accuracy, we report Precision, F1 score, and Matthews Correlation Coefficient to capture changes in both overall and class-sensitive reliability. The analysis shows that the main performance loss occurs when models are tested at noise levels higher than those encountered during training, whereas degradation is less disruptive when the noise distribution is consistent across both phases. These results underscore the importance of explicitly accounting for noise variability when deploying a skin lesion classifier in real-world clinical settings.

: The FORUM (Far-infrared Outgoing Radiation Understanding and Monitoring) mission will provide, for the first time, systematic far-infrared spectral measurements of Earth's outgoing radiation, enabling improved understanding of atmospheric processes and the radiation budget. Retrieving atmospheric states from these observations constitutes a high-dimensional, ill-posed inverse problem, particularly under cloudy-sky conditions where multiple-scattering effects are present. In this work, we develop a data-driven, physics-aware inversion framework for FORUM all-sky retrievals based on latent twins: coupled autoencoders for atmospheric states and spectra, combined with bidirectional latent-space mappings. A lightweight model-consistency correction ensures physically plausible cloud variable reconstructions. The resulting framework demonstrates potential for retrievals of atmospheric, cloud and surface variables, providing information that can serve as a prior, initial guess, or surrogate for computationally expensive full-physics inversion methods. It also enables robust scene classification and near-instantaneous inference, making it suitable for operational near-real-time applications. We demonstrate its performance on synthetic FORUM-like data and discuss implications for future data assimilation and climate studies.

In the forward model for limb-scanning instruments, ray tracing must be accounted for because variations in air refractivity cause the lines of sight to bend from straight paths into curves. The tangent point of a line of sight defined as the minimum height, depends on both the instrument’s nadir angle and the atmospheric state. To achieve reliable tangent point determination, the off-nadir angles must be calibrated to account for Earth’s ellipsoidal geometry and realistic atmospheric conditions sampled by these lines of sight paths. In this study, we improve the ray-tracing algorithm originally developed for MIPAS operational retrievals. The algorithm is applied to more accurately assess the impact of atmospheric variability on the relationship between off-nadir angles and tangent point localization. It also allows us to estimate the smoothing error associated with the coarse horizontal sampling of the MIPAS observation pattern. We used a configuration representative of the proposed CAIRT mission instrument as a case scenario for future instruments such as the STRIVE mission by NASA.

Invasive alien species pose major ecological and socio-economic challenges, especially when they exert strong predation pressure on native resources while offering opportunities for commercial use. Addressing this dual role requires models that integrate ecological dynamics, economic incentives, and institutional policies. This paper develops a general bioeconomic framework for a predator–prey system subject to human interventions. The native resource is harvested for economic purposes, while the invasive predator can be controlled through both commercial exploitation and institutional removal. The analysis combines equilibrium methods and optimal control over finite horizons, capturing the interaction between long-run ecological regimes and short-term management objectives. The results characterize threshold conditions for species coexistence and identify how harvesting and institutional removal affect the ecological regimes of the system. The analysis shows that moderate exploitation of the invasive predator can enlarge the coexistence region, whereas eradication requires removal efforts exceeding its intrinsic growth rate. The optimal control analysis highlights the role of coordinated management instruments, with numerical results indicating that mixed market and institutional strategies outperform single-instrument policies. The framework is inspired by the interaction between the invasive blue crab (Callinectes sapidus) and native bivalves in Mediterranean coastal systems, and points to policy strategies that integrate ecological conservation, economic valorization, and institutional containment.

This paper addresses the numerical modeling of isentropic gas dynamics on one-dimensional networks, focusing on the Euler equations with a novel class of transmission conditions at network junctions, termed the Jump Transmission Condition (JTC). Unlike traditional models that enforce continuity of density at junctions, the (JTC) allows for a density jump proportional to the flux, reflecting phenomena such as bottlenecks in biological or traffic networks. The authors propose a relaxation scheme based on the vector BGK approach, which ensures mass conservation and enforces the (JTC) without requiring the explicit solution of the Riemann problem at the junction. The scheme is analyzed for its mathematical properties, including entropy dissipation and positivity, and is compared with classical solvers based on the explicit solution of the Riemann problem. Numerical experiments on simple and complex networks demonstrate the accuracy, robustness, and flexibility of the proposed method, especially in handling both subsonic and supersonic regimes and in extending to general networks.

Il presente report ricostruisce l'evoluzione della valutazione della terza missione nell'ambito degli esercizi VQR condotti da ANVUR tra il 2004 e il 2024, analizzandone criticamente metodologia, processi e ricadute sul sistema della ricerca pubblica italiana. Attraverso un'applicazione del modello Business Process Management (BPM), si valuta se i processi valutativi si siano effettivamente semplificati e ottimizzati nel tempo, evidenziando persistenti criticità in termini di tempi, qualità dei dati e normalizzazione rispetto al contesto istituzionale. Mediante un'analisi Analytic Hierarchy Process (AHP), il lavoro verifica inoltre l'ipotesi di un modello valutativo "università-centrico", individuando aree di indicatori significativamente sbilanciate a favore degli atenei — in particolare formazione continua e gestione di beni culturali — a fronte di una sostanziale neutralità degli indicatori di trasferimento tecnologico. Infine, il report esamina il legame tra risultati della valutazione e meccanismi di finanziamento, rilevando un progressivo rafforzamento della componente premiale per le università e, al contrario, la scomparsa di ogni riferimento esplicito alla VQR nei criteri di riparto del FOE per gli enti pubblici di ricerca dal 2016 in poi. Il lavoro si chiude con alcune proposte di ricalibrazione del sistema valutativo.

We introduce VCKNet (Variable-sized Convolutional Kernel Network), an adaptive, modular, and lightweight Convolutional Neural Network (CNN) with three kernel-scale branches, channel attention, and dynamic fusion. These components recalibrate scale-aware features to reduce redundancy and support discriminative learning. The proposed architectural design preserves computational efficiency while maintaining a clear structural organization. Experiments on standard benchmarks and in-the-wild datasets show that VCKNet effectively captures multiscale structure, achieving performance competitive with that of deeper state-of-the-art models. Statistical analysis across repeated runs further confirms VCKNet’s stability. Computational cost and deployment efficiency analyses further support its suitability for resource-constrained and production-oriented scenarios where flexibility, stability, and conceptual clarity are essential.

We consider a variant of the sticky disk energy where distances between particles are evaluated through the sup norm ‖⋅‖∞in the plane. We first prove crystallization of minimizers in the square lattice, for any fixed number N of particles. Then we consider the limit as N →∞: In contrast to the standard sticky disk, there is only one orientation in the limit, and we are able to compute explicitly the Γ-limit to be an anisotropic perimeter with octagonal Wulff shape. The results are based on an energy decomposition for graphs that generalizes the one proved by De Luca-Friesecke [10] in the triangular case.

This paper deals with the dynamics-driven by the gradient flow of negative fractional seminorms-of empirical measures towards equi-spaced ground states. Specifically, we consider periodic empirical measures μ on the real line that are screened by the Lebesgue measure, i.e., with μ-d x μ-dx having zero average. To each of these measures μ we associate a (periodic) function u satisfying u ′ = d x-μ {u′=dx-μ. For s (0, 1 2) {e(0, 1{2) we introduce energy functionals s (μ) Es(μ) that can be understood as the density of the s-Gagliardo seminorm of u per unit length. Since for s ≥ 1 2 {s1/2, the s-Gagliardo seminorms are infinite on functions with jumps, some regularization procedure is needed: For s [ 1 2, 1) e1{2,1) we define ε s (μ):= s (μ ε) Es(μ):= Es(μ), where μ ε μ is obtained by mollifying μ on scale ε. We prove that the minimizers of s Es and ε s Es are the equi-spaced configurations of particles with lattice spacing equal to one. Then we prove the exponential convergence of the corresponding gradient flows to the equi-spaced steady states. Finally, although for s [ 1 2, 1) e1/2,1) the energy functionals ε s Es blow up as ε → 0 to 0, their gradients are uniformly bounded (with respect to ε), so that the corresponding trajectories converge, as ε → 0 to 0, to the gradient flow solution of a suitable renormalized energy.

In this paper, a multidisciplinary design optimization algorithm, the Normal Boundary Intersection (NBI) method, is applied to the design of some devices of a sailing yacht. The full Pareto front is identified for two different design problems, and the optimal configurations are compared with standard devices. The great efficiency of the optimization algorithm is demonstrated by the wideness and density of the identified Pareto front.

We provide explicit upper bounds on some distances between the (law of the) output of a random Gaussian neural network and (the law of) a random Gaussian vector. Our main results concern deep random Gaussian neural networks, with a rather general activation function. The upper bounds show how the widths of the layers, the activation function and other architecture parameters affect the Gaussian approximation of the output. Our techniques, relying on Stein's method and integration by parts formulas for the Gaussian law, yield estimates on distances which are indeed integral probability metrics, and include the convex distance. This latter metric is defined by testing against indicator functions of measurable convex sets, and so allows for accurate estimates of the probability that the output is localized in some region of the space. Such estimates have a significant interest both from a practitioner's and a theorist's perspective.

In this manuscript, we propose a stable algorithm for computing the zeros of Althammer polynomials. These polynomials are orthogonal with respect to a Sobolev inner product, and are even if their degree is even, odd otherwise. Furthermore, their zeros are real, distinct, and located inside the interval (−1, 1). The Althammer polynomial p_n(x) of degree n satisfies a long recurrence relation, whose coefficients can be arranged into a Hessenberg matrix of order n, with eigenvalues equal to the zeros of the considered polynomial. Unfortunately, the eigenvalues of this Hessenberg matrix are very ill–conditioned, and standard balancing procedures do not improve their condition numbers. Here, we introduce a novel algorithm for computing the zeros of p_n(x), which first transforms the Hessenberg matrix into a similar symmetric tridiagonal one, i.e., a matrix whose eigenvalues are perfectly conditioned, and then computes the zeros of p_n(x) as the eigenvalues of the latter tridiagonal matrix. Moreover, we propose a second algorithm, faster but less accurate than the former one, which computes the zeros of p_n(x) as the eigenvalues of a truncated Hessenberg matrix, obtained by properly neglecting some diagonals in the upper part of the original matrix. The computational complexity of the proposed algorithms are, respectively, O( n^3/ 6 ), and O(l^2 n), with l<

In this paper, we derive a new method to compute the nodes and weights of simultaneous n-point Gaussian quadrature rules. The method is based on the eigendecomposition of the banded lower Hessenberg matrix that contains the coefficients of the recurrence relations for the corresponding multiple orthogonal polynomials. The novelty of the approach is that it uses the property of total nonnegativity of this matrix associated with the particular considered multiple orthogonal polynomials, in order to compute its eigenvalues and eigenvectors in a numerically stable manner. The overall complexity of the computation of all the nodes and weights is O(n^2).

Over the last decade, the Lattice Boltzmann method has found major scope for the simulation of a large spectrum of problems in soft matter, from multiphase and multi-component microfluidic flows, to foams, emulsions, colloidal flows, to name but a few. Crucial to many such applications is the role of supramolecular interactions which occur whenever mesoscale structures, such as bubbles or droplets, come in close contact, say of the order of tens of nanometers. Regardless of their specific physico-chemical origin, such near-contact interactions are vital to preserve the coherence of the mesoscale structures against coalescence phenomena promoted by capillarity and surface tension, hence the need of including them in Lattice Boltzmann schemes. Strictly speaking, this entails a complex multiscale problem, covering about six spatial decades, from centimeters down to tens of nanometers, and almost twice as many in time. Such a multiscale problem can hardly be taken by a single computational method, hence the need for coarse-grained models for the near-contact interactions. In this review, we shall discuss such coarse-grained models and illustrate their application to a variety of soft flowing matter problems, such as soft flowing crystals, strongly confined dense emulsions, flowing hierarchical emulsions, soft granular flows, as well as the transmigration of active droplets across constrictions. Finally, we conclude with a few considerations on future developments in the direction of quantum-nanofluidics, machine learning, and quantum computing for soft flows applications.

In this paper, we explore the determination of a spectral emissivity profile that closely matches real data, intended for use as an initial guess and/or a priori information in a retrieval code. Our approach employs a Bayesian method that integrates the CAMEL (Combined ASTER MODIS Emissivity over Land) emissivity database with the MODIS/Terra+Aqua Yearly Land Cover Type database. The solution is derived as a convex combination of high-resolution Huang profiles using the Bayesian framework. We test our method on IASI (Infrared Atmospheric Sounding Interferometer) data and find that it outperforms the linear spline interpolation of the CAMEL data and the Huang emissivity database itself.

An X-ray diffraction pattern consists of relevant information (the signal) and noisy background. Under the assumption that they behave as the components of a two-dimensional mixture (bicomponent fluid) having slightly different physical properties related to the density gradients, a Lattice Boltzmann Method is applied to disentangle the two different diffusive dynamics. The solution is numerically stable, not computationally demanding, and, it also provides an efficient increase in the signal-to-noise ratio for patterns blurred by Poissonian noise and affected by collection data anomalies (fiber-like samples, experimental setup, etc.). The model is succesfully applied to different resolution images.

Mechanotransduction is the process that enables the conversion of mechanical cues into biochemical signaling. While all our cells are well known to be sensitive to such stimuli, the details of the systemic interaction between mechanical input and inflammation are not well integrated. Often, indeed, they are considered and studied in relatively compartmentalized areas, and we therefore argue here that to understand the relationship of mechanical stimuli with inflammation – with a high translational potential - it is crucial to offer and analyze a unified view of mechanotransduction. We therefore present here pathway representation, recollected with the standard systems biology markup language (SBML) and explored with network biology approaches, offering RAC1 as an exemplar and emerging molecule with potential for medical translation.

A reaction–diffusion system governing the predator–prey interaction with specialist predator and herd behavior for prey is investigated. Linear stability of the interior equilibrium is studied, and conditions guaranteeing the occurrence of Turing instability, induced by cross-diffusion, are found, with a full characterization of the Turing instability region in the parameter space. Numerical simulations on the obtained results are provided.

On the half line, we introduce a new sequence of near-best uniform approximation polynomials, easily computable by the values of the approximated function at a truncated number of Laguerre zeros. Such approximation polynomials come from a discretization of filtered Fourier–Laguerre partial sums, which are filtered using a de la Vallée Poussin (VP) filter. They have the peculiarity of depending on two parameters: a truncation parameter that determines how many of the n Laguerre zeros are considered, and a localization parameter, which determines the range of action of the VP filter we will apply. As n→∞, under simple assumptions on such parameters and the Laguerre exponents of the involved weights, we prove that the new VP filtered approximation polynomials have uniformly bounded Lebesgue constants and uniformly convergence at a near–best approximation rate, for any locally continuous function on the semiaxis. The numerical experiments have validated the theoretical results. In particular, they show a better performance of the proposed VP filtered approximation versus the truncated Lagrange interpolation at the same nodes, especially for functions a.e. very smooth with isolated singularities. In such cases, we see a more localized approximation and a satisfactory reduction of the Gibbs phenomenon.