One of the milestones to be achieved in the design of the bolometric diagnostics for the new Italian Divertor Tokamak Test (DTT) project is the estimation of the radiated power at the start of operation, i.e. the so-called first plasma, in order to perform various tasks ranging from scientific analysis and planning of the discharges to the feedback protection of the machine. In fact, real-time (RT) feedback control of the radiation pattern for prevention is both a delicate and important matter, for example in terms of mitigating and avoiding disruptions. It would therefore be desirable to monitor not only the total power emitted, but also the one emitted by the different regions of the plasma. This paper then focuses on showing the initial design of the main strategy for estimating the plasma radiation in two different situations: for RT control and for an inter-shot analysis. The first approach for RT then, is based on the estimation of the radiated power inside the first wall using specific lines of sight (LoS). Such estimates have been compared with those obtained from slower tomographic reconstructions of synthetic emissivity profiles (phantoms). Furthermore, a first design of the Region Of Interest (ROI) for a fast implementation of an already established macro-estimation of the radiated power in different locations of the main chamber is provided and the overall method is adapted for DTT. Regarding the design of the inter-shot data exploitation then, since tomographic reconstructions will most likely be available during an inter-shot basis, it is planned to provide a more accurate estimate of the radiated power from different locations of the device for a better design and tuning of the discharges. In order to achieve such a long-term goal, an initial strategy for adapting a maximum likelihood based algorithm for inter-shot analysis is described.

On both the Joint European Torus (JET) and the ASDEX-Upgrade (AUG) tokamaks, an Expectation Maximization algorithm has been adapted to implement a Maximum Likelihood (ML) approach to derive tomograms from bolometers data. The main feature of such an approach is the ability to estimate the variance associated with the reconstructed tomograms and hence the uncertainties in the derived quantities. It has therefore been selected to support the design of the bolometric diagnostics for the Divertor Tokamak Test Facility (DTT). A reliable reconstruction of the emissivity profile is indeed relevant for the scientific exploitation of the device. As any tomographic inversion method is an ill-conditioned problem, two milestones should be achieved: ensuring the reconstruction of specific features and minimizing the risk of producing artefacts. The design strategy for the bolometric diagnostic on DTT involves testing various possible layouts to ensure that they meet the above requirements and are compatible with the engineering and machine constraints. This contribution focuses on showing how the conceptual bolometric layout can handle different phantoms mimicking typical emissivities observed on JET and AUG. A methodology has been developed to further optimize the layout within the constraints of the machine design. For the initial phase of DTT, a reduced layout is proposed, utilizing half of the conceptual lines of sight. The ability to reconstruct specific emissivity features while minimizing the risk of producing artefacts has been tested. The current ML implementation uses an anisotropic diffusion technique and is already significantly faster than the JET implementation for each reconstruction.

In the European roadmap towards nuclear fusion, the new Italian project DTT (Divertor Tokamak Test), currently under construction at the ENEA Frascati Research Centre, aims at exploring alternative solutions for the divertor, in support to the optimization of the divertor configuration foreseen for DEMO. A dedicated set of diagnostics are planned to measure the total radiated power and to monitor plasma processes by means of Soft-X emission measurements. The total radiated power acquisition will rely on commercial metal resistor bolometers, while for the SXR range of energies a new technology based on custom Chemical Vapor Deposition diamonds (CVD) will be adopted. This work focuses on the mechanical layout design of these fundamental diagnostics, respecting geometrical and functional constrains of DTT and minimizing the diagnostics volume inside the access pipe. An integrated and compact solution, allowing flexible positioning and easy maintenance of the two systems has been identified, with a structural layout based on a modular approach. The detectors of the two diagnostics are installed on a common frame, with adjustable mounts for an independent fine-tuning of their alignment. The high heat load from the plasma is coped with by means of a custom active water-cooling system, to protect the sensors while ensuring stable and reliable operations. To allow a good physical exploitation of the collected data, i.e. with tomographic reconstructions, the lines of sight of both systems must be properly arranged on the poloidal section of the plasma. The proposed mechanical design foresees a customized chassis for each of the four poloidal ports of the DTT vacuum vessel, assuring a suitable coverage of the plasma's poloidal cross-section.

Background Analysis of Electronic Health Records (EHRs) is crucial in real-world evidence (RWE), especially in oncology, as it provides valuable insights into the complex nature of the disease. The implementation of advanced techniques for automated extraction of structured information from textual data potentially enables access to expert knowledge in highly specialized contexts. In this paper, we introduce MISTIC, a Natural Language Processing (NLP) approach to classify the presence or absence of metastasis in Italian EHRs, in the breast cancer domain. Methods Our approach consists of a transformer-based framework designed for few-shot learning, requiring a small labelled dataset and minimal computational resources for training. The pipeline includes text segmentation to improve model processing and topic analysis to filter informative content, ensuring relevant input data for classification. Results MISTIC was evaluated across multiple data sources, and compared to several benchmark methodologies, ranging from a pattern-matching system, composed of regex and semantic rules, to BERT-based models implemented in a zero-shot learning setup and Large Language Models (LLMs). The results demonstrate the generalization of our approach, achieving an F-Score above 87% on all the sources, and outperforming the other experiments, with an overall F-Score of 91.2%. Conclusions MISTIC achieves high performance in the Italian metastasis classification task, outperforming rule-based systems, zero-shot BERT models, and LLMs. Its few-shot learning setup offers a computationally efficient alternative to large-scale models, while its segmentation and topic analysis steps enhance explainability by explicitly linking predictions to key textual elements. Furthermore, MISTIC demonstrates strong generalization across different data sources, reinforcing its potential as a scalable and transparent solution for clinical text classification. By extracting high-quality metastatic information from diverse textual data, MISTIC supports medical researchers in analyzing unstructured and highly informative content across a wide range of medical reports. In doing so, it enhances data accessibility and interpretability, addressing a critical gap in health informatics and clinical practice.

We

The increasing availability of clinical reports offers valuable opportunities for natural language processing (NLP) applications in healthcare. Large Language Models (LLMs), such as BERT-based architectures and generative models, have shown great promise in text classification, summarization, and semantic analysis. However, applying LLMs to Electronic Health Records (EHRs) poses challenges due to token limits and the complexity of clinical text. Sentence Boundary Detection (SBD), which segments text into meaningful units, is a critical preprocessing step to address token constraints and improve model interpretability, particularly for tasks like text classification. This study benchmarks several SBD methods, including traditional approaches (e.g., NLTK, Stanza, PySBD) and state-of-the-art transformer-based models, such as Segment Any Text (SAT), fine-tuned using low-rank adaptation (LoRA) for the clinical domain. The models were evaluated on a dataset of clinical reports in Italian, sourced from the Gemelli hospital of Rome, using metrics like F1-score to measure segmentation quality. The results reveal that PySBD achieved the best performance, closely aligning with the gold standard, with a median F1-Score of 83%. We also assessed the impact of segmentation on a downstream metastasis classification task, comparing the performance of a transformer-based model applied to unsegmented reports versus reports processed with PySBD. Segmentation outperformed the entire report scenario, with a higher F1-Score of 92% versus 88%, demonstrating that SBD improves text classification by ensuring semantic coherence, adhering to token constraints, and providing sentence-level explainability. In conclusion, this study highlights the importance of SBD in enhancing both the quality and interpretability of downstream NLP tasks in healthcare. By benchmarking traditional and transformer-based SBD models, we validate the role of segmentation as a critical preprocessing step to advance clinical NLP applications, offering insights for improving performance and clinical relevance in the processing of EHRs.

We give analytic description for the completion of C?0 (R+) in Dirichletspace D1,p(R+, ?) := {u : R+ -> R : u is locally absolutely continuous on R+ and ||u? ||_Lp(R+,?) < ?}, for given continuouspositive weight ? defined on R+, where 1 < p < ?. The conditions are described in terms of the modified variants of the Bpconditions due to Kufner and Opic from 1984, which in our approach are focusing on integrability of ?^-p/(p-1) near zero or near infinity. Moreover, we propose applications of our results to: obtaining newvariants of Hardy inequality, interpretation of boundary value problems in ODE's defined on the halpfline with solutions in D1,p(R+, ?),new results from complex interpolation theory dealing with interpolation spaces between weighted Dirichlet spaces, and to derivationof new Morrey type embedding theorems for our Dirichlet space.

Planning petrol station replenishment is an important logistics activity for all the major oil companies. The studied Multi-Depot Periodic Petrol Station Replenishment problem derives from a real case in which the company must replenish a set of petrol stations from a set of depots, during a weekly planning horizon. The company must ensure refuelling according to available visiting patterns, which can be different from customer to customer. A visiting pattern predefines how many times (days) the replenishment occurs during a week and in which visiting days a certain amount of fuel must be delivered. To fulfill the weekly demand of each petrol station, one of the available replenishment plans must be selected among a given set of visiting patterns. The aim is to minimize the total distance travelled by the fleet of tank trucks during the entire planning horizon. A matheuristic approach is proposed, based on the cluster-first route-second paradigm, to solve it. The proposed approach is thoroughly tested on a set of realistic random instances. Finally, a weekly large real instance is considered with 194 petrol stations and two depots.

Magnetoencephalography (MEG) is a valuable non-invasive neurophysiology technique for investigation of brain function and dysfunction. In this chapter, we will discuss the main characteristics of MEG signals, and the great potential it offers for scientific interrogation in psychology, cognitive neuroscience, neurology, and neuropsychiatry. Starting from the physical properties of MEG recordings, the chapter will highlight the main advantages of utilizing MEG in neuroscience (that is a combination of very high temporal resolution and good spatial resolution) and will summarize the current status of MEG in research and clinical settings. To make this topic more relatable to widely available electroencephalography (EEG), we will present several comparisons of MEG with EEG. The objective of the present chapter is to provide a broad overview of the principle concepts and strengths of MEG, aimed at newcomers to the field.

We prove the nonlinear asymptotic stability of stably stratified solutions to the Incompressible Porous Media equation (IPM) for initial perturbations in $\dot H^{1-\tau}(\R^2) \cap \dot H^s(\R^2)$ with $s > 3$ and for any $0 < \tau <1$. Such result improves the existing literature, where the asymptotic stability is proved for initial perturbations belonging at least to $H^{20}(\R^2)$. \\ More precisely, the aim of the article is threefold. First, we provide a simplified and improved proof of global-in-time well-posedness of the Boussinesq equations with strongly damped vorticity in $H^{1-\tau}(\R^2) \cap \dot H^s(\R^2)$ with $s > 3$ and $0 < \tau <1$. Next, we prove the strong convergence of the Boussinesq system with damped vorticity towards (IPM) under a suitable scaling. Lastly, the asymptotic stability of stratified solutions to (IPM) follows as a byproduct.\\ A symmetrization of the approximating system and a careful study of the anisotropic properties of the equations via anisotropic Littlewood-Paley decomposition play key roles to obtain uniform energy estimates. Finally, one of the main new and crucial points is the integrable time decay of the vertical velocity $\|u_2(t)\|_{L^\infty (\R^2)}$ for initial data only in $\dot H^{1-\tau}(\R^2) \cap \dot H^s(\R^2)$ with $s >3$.

This article is concerned with the rigorous justification of the hydrostatic limit for continuouslystratified incompressible fluids under the influence of gravity.The main peculiarity of this work with respect to previous studies is that no (regularizing) viscosity contributionis added to the fluid-dynamics equations and only diffusivity effects are included. Motivated byapplications to oceanography, the diffusivity effects included in this work are induced by an advection termwhose specific form was proposed by Gent and McWilliams in the 90's to model effective eddy correlations fornon-eddy-resolving systems.The results of this paper heavily rely on the assumption of stable stratification. We provide the wellposednessof the hydrostatic equations and of the original (non-hydrostatic) equations for stably stratified fluids,as well as their convergence in the limit of vanishing shallow-water parameter. The results are established inhigh but finite Sobolev regularity and keep track of the various parameters at stake.A key ingredient of our analysis is the reformulation of the systems by means of isopycnal coordinates,which allows to provide careful energy estimates that are far from being evident in the original coordinatesystem.

We prove the strong ill-posedness of the two-dimensional Boussinesq system in vorticity form in L8pR2qwithout boundary, building upon the method that Shikh Khalil & Elgindi arXiv:2207.04556v1 developed for scalarequations. We provide examples of initial data with vorticity and density gradient of small L8pR2q size, for which thehorizontal density gradient has a strong L8pR2q-norm inflation in infinitesimal time, while the vorticity and the verticaldensity gradient remain bounded. Furthermore, exploiting the three-dimensional version of Elgindi's decomposition ofthe Biot-Savart law, we apply our method to the three-dimensional axisymmetric Euler equations with swirl and awayfrom the vertical axis, showing that a large class of initial data with vorticity uniformly bounded and small in L8pR2qprovides a solution whose gradient of the swirl has a strong L8pR2q-norm inflation in infinitesimal time. The norminflations are quantified from below by an explicit lower bound which depends on time, the size of the data and is validfor small times

It is known that executing a perfect shifted QR step via the implicit QR algorithm may not result in a deflation of the perfect shift. Typically, several steps are required before deflation actually takes place. This deficiency can be remedied by determining the similarity transformation via the associated eigenvector. Similar techniques have been deduced for the QZ algorithm and for the rational QZ algorithm. In this paper we present a similar approach for executing a perfect shifted QZ step on a general rank structured pencil instead of a specific rank structured one, e.g., a Hessenberg--Hessenberg pencil. For this, we rely on the rank structures present in the transformed matrices. A theoretical framework is presented for dealing with general rank structured \rev{pencils} and deflating subspaces. We present the corresponding algorithm allowing} to deflate simultaneously a block of eigenvalues rather than a single one. We define the level-rho poles and show that these poles are maintained executing the deflating algorithm. Numerical experiments illustrate the robustness of the presented approach showing the importance of using the improved scaled residual approach.

In this paper, we develop new methods to join machine learning techniques and macroscopic differential models, aimed at estimate and forecast vehicular traffic. This is done to complement respective advantages of data-driven and model-driven approaches. We consider here a dataset with flux and velocity data of vehicles moving on a highway, collected by fixed sensors and classified by lane and by class of vehicle. By means of a machine learning model based on an LSTM recursive neural network, we extrapolate two important pieces of information: (1) if congestion is appearing under the sensor, and (2) the total amount of vehicles which is going to pass under the sensor in the next future (30 min). These pieces of information are then used to improve the accuracy of an LWR-based first-order multi-class model describing the dynamics of traffic flow between sensors. The first piece of information is used to invert the (concave) fundamental diagram, thus recovering the density of vehicles from the flux data, and then inject directly the density datum in the model. This allows one to better approximate the dynamics between sensors, especially if an accident/bottleneck happens in a not monitored stretch of the road. The second piece of information is used instead as boundary conditions for the equations underlying the traffic model, to better predict the total amount of vehicles on the road at any future time. Some examples motivated by real scenarios will be discussed. Real data are provided by the Italian motorway company Autovie Venete S.p.A.

The ongoing reproducibility crisis in psychology and cognitive neuroscience has sparked increasing calls to re-evaluate and reshape scientific culture and practices. Heeding those calls, we have recently launched the EEGManyPipelines project as a means to assess the robustness of EEG research in naturalistic conditions and experiment with an alternative model of conducting scientific research. One hundred sixty-eight analyst teams, encompassing 396 individual researchers from 37 countries, independently analyzed the same unpublished, representative EEG data set to test the same set of predefined hypotheses and then provided their analysis pipelines and reported outcomes. Here, we lay out how large-scale scientific projects can be set up in a grassroots, community-driven manner without a central organizing laboratory. We explain our recruitment strategy, our guidance for analysts, the eventual outputs of this project, and how it might have a lasting impact on the field.

Two different direct methods are proposed to solve Cauchy singular integral equations on the real line. The aforementioned methods differ in order to be able to prove their convergence which depends on the smoothness of the known term function in the integral equation.

The dynamics of stabilised concentrated emulsions presents a rich phenomenology including chaotic emulsification, non-Newtonian rheology and ageing dynamics at rest. Macroscopic rheology results from the complex droplet microdynamics and, in turn, droplet dynamics is influenced by macroscopic flows via the competing action of hydrodynamic and interfacial stresses, giving rise to a complex tangle of elastoplastic effects, diffusion, breakups and coalescence events. This tight multiscale coupling, together with the daunting challenge of experimentally investigating droplets under flow, has hindered the understanding of concentrated emulsions dynamics. We present results from three-dimensional numerical simulations of emulsions that resolve the shape and dynamics of individual droplets, along with the macroscopic flows. We investigate droplet dispersion statistics, measuring probability density functions (p.d.f.s) of droplet displacements and velocities, changing the concentration, in the stirred and ageing regimes. We provide the first measurements, in concentrated emulsions, of the relative droplet–droplet separations p.d.f. and of the droplet acceleration p.d.f., which becomes strongly non-Gaussian as the volume fraction is increased above the jamming point. Cooperative effects, arising when droplets are in contact, are argued to be responsible of the anomalous superdiffusive behaviour of the mean square displacement and of the pair separation at long times, in both the stirred and in the ageing regimes. This superdiffusive behaviour is reflected in a non-Gaussian pair separation p.d.f., whose analytical form is investigated, in the ageing regime, by means of theoretical arguments. This work paves the way to developing a connection between Lagrangian dynamics and rheology in concentrated emulsions.

The paper deals with the numerical solution of Cauchy Singular Integral Equations based on some non standard polynomial quasi-projection of de la Vallée Poussin type. Such kind of approximation presents several advantages over classical Lagrange interpolation such as the uniform boundedness of the Lebesgue constants, the near-best order of uniform convergence to any continuous function, and a strong reduction of Gibbs phenomenon. These features will be inherited by the proposed numerical method which is stable and convergent, and provides a near-best polynomial approximation of the sought solution by solving a well conditioned linear system. The numerical tests confirm the theoretical error estimates and, in case of functions subject to Gibbs phenomenon, they show a better local approximation compared with analogous Lagrange projection methods.

We consider here a cell-centered finite difference approximation of the Richards equation in three dimensions, averaging for interface values the hydraulic conductivity, a highly nonlinear function, by arithmetic, upstream and harmonic means. The nonlinearities in the equation can lead to changes in soil conductivity over several orders of magnitude and discretizations with respect to space variables often produce stiff systems of differential equations. A fully implicit time discretization is provided by backward Euler one-step formula; the resulting nonlinear algebraic system is solved by an inexact Newton Armijo-Goldstein algorithm, requiring the solution of a sequence of linear systems involving Jacobian matrices. We prove some new results concerning the distribution of the Jacobians eigenvalues and the explicit expression of their entries. Moreover, we explore some connections between the saturation of the soil and the ill conditioning of the Jacobians. The information on eigenvalues justifies the effectiveness of some preconditioner approaches which are widely used in the solution of Richards equation. We also propose a new software framework to experiment with scalable and robust preconditioners suitable for efficient parallel simulations at very large scales. Performance results on a literature test case show that our framework is very promising in the advance toward realistic simulations at extreme scale.

In this manuscript we propose a numerical method for non-linear integro-differential systems arising in age-of-infection models in a heterogeneously mixed population. The discrete scheme is based on direct quadrature methods and provides an unconditionally positive and bounded solution. Furthermore, we prove the existence of the numerical final size of the epidemic and show that it tends to its continuous equivalent as the discretization steplength vanishes.