arXiv subjects

Luca Barbieri

Publications and source records attributed to Luca Barbieri.

27 records · Page 2Linked to original sources

arXiv2024

A Secure and Trustworthy Network Architecture for Federated Learning Healthcare Applications

Federated Learning (FL) has emerged as a promising approach for privacy-preserving machine learning, particularly in sensitive domains such as healthcare. In this context, the TRUSTroke project aims to leverage FL to assist clinicians in ischemic stroke prediction. This paper provides an overview of the TRUSTroke FL network infrastructure. The proposed architecture adopts a client-server model with a central Parameter Server (PS). We introduce a Docker-based design for the client nodes, offering a flexible solution for implementing FL processes in clinical settings. The impact of different communication protocols (HTTP or MQTT) on FL network operation is analyzed, with MQTT selected for its suitability in FL scenarios. A control plane to support the main operations required by FL processes is also proposed. The paper concludes with an analysis of security aspects of the FL architecture, addressing potential threats and proposing mitigation strategies to increase the trustworthiness level.

cs.AI↗

arXiv2024

Deep Learning-based Cooperative LiDAR Sensing for Improved Vehicle Positioning

Accurate positioning is known to be a fundamental requirement for the deployment of Connected Automated Vehicles (CAVs). To meet this need, a new emerging trend is represented by cooperative methods where vehicles fuse information coming from navigation and imaging sensors via Vehicle-to-Everything (V2X) communications for joint positioning and environmental perception. In line with this trend, this paper proposes a novel data-driven cooperative sensing framework, termed Cooperative LiDAR Sensing with Message Passing Neural Network (CLS-MPNN), where spatially-distributed vehicles collaborate in perceiving the environment via LiDAR sensors. Vehicles process their LiDAR point clouds using a Deep Neural Network (DNN), namely a 3D object detector, to identify and localize possible static objects present in the driving environment. Data are then aggregated by a centralized infrastructure that performs Data Association (DA) using a Message Passing Neural Network (MPNN) and runs the Implicit Cooperative Positioning (ICP) algorithm. The proposed approach is evaluated using two realistic driving scenarios generated by a high-fidelity automated driving simulator. The results show that CLS-MPNN outperforms a conventional non-cooperative localization algorithm based on Global Navigation Satellite System (GNSS) and a state-of-the-art cooperative Simultaneous Localization and Mapping (SLAM) method while approaching the performances of an oracle system with ideal sensing and perfect association.

eess.SP↗

arXiv2023

Temperature inversion in a gravitationally bound plasma: Case of the solar corona

The temperature of the solar atmosphere increases from thousands to millions of degrees moving from the lower layer (chromosphere) to the outermost one (corona), while the density drops accordingly. The mechanism behind this phenomenon, known as a temperature inversion, is still unknown. In this work, we model a coronal loop as a collisionless plasma confined in a semicircular tube that is subject to the Sun's gravity and in thermal contact with a fully collisional chromosphere behaving as a thermostat at the loop's feet. By using kinetic $N$-particle simulations and analytical calculations, we show that rapid, intermittent, and short-lived heating events in the chromosphere drive the coronal plasma towards a non-equilibrium stationary state. The latter is characterized by suprathermal tails in the particles' velocity distribution functions, exhibiting temperature and density profiles strikingly similar to those observed in the atmosphere of the Sun. These results suggest that a million-Kelvin solar corona can be produced without the local deposition of heat in the upper layer of the atmosphere that is typically assumed by standard approaches. We find that suprathermal distribution functions in the corona are self-consistently produced instead of postulated a priori, in contrast to classical kinetic models based on a velocity filtration mechanism.

astro-ph.SR↗

arXiv2023

Fast emitting nanocomposites for high-resolution ToF-PET imaging based on multicomponent scintillators

Time-of-Flight Positron Emission Tomography is a medical imaging technique, based on the detection of two back-to-back γ-photons generated from radiotracers injected in the body. Its limit is the ability of employed scintillation detectors to discriminate in time the arrival of γ-pairs, i.e. the coincidence time resolution (CTR). A CTR < 50 ps that would enable fast imaging with ultralow radiotracer dose. Monolithic materials do not have simultaneously the required high light output and fast emission characteristics, thus the concept of scintillating heterostructure is proposed, where the device is made of a dense scintillator coupled to a fast-emitting light material. Here we present a composite polymeric scintillator, whose density has been increased upon addition of hafnium oxide nanoparticles. This enhanced by +300% its scintillation yield, surpassing commercial plastic scintillators. The nanocomposite is coupled to bismuth germanate oxide (BGO) realizing a multilayer scintillator. We observed the energy sharing between its components, which activate the nanocomposite fast emission enabling a net CTR improvement of 25% with respect to monolithic BGO. These results demonstrate that a controlled loading with dense nanomaterials is an excellent strategy to enhance the performance of polymeric scintillators for their use in advanced radiation detection and imaging technologies.

physics.ins-det↗

arXiv2023

Dynamics of intermediate mass black holes in globular clusters. Wander radius and anisotropy profiles

We recently introduced a new method for simulating collisional gravitational N-body systems with approximately linear time scaling with $N$, based on the Multi-Particle Collision (MPC) scheme, previously applied in Plasma Physics. We simulate globular clusters with a realistic number of stellar particles (at least up to several times $10^6$) on a standard workstation. We simulate clusters hosting an intermediate mass black hole (IMBH), probing a broad range of BH-cluster and BH-average-star mass ratios, unrestricted by the computational constraints affecting direct N-body codes. We use either single mass models or models with a Salpeter mass function, with the IMBH initially sitting at the centre. The force exerted by and on the IMBH is evaluated with a direct scheme. We measure the evolution of the Lagrangian radii and core density and velocity dispersion over time. In addition, we study the evolution of the velocity anisotropy profiles. We find that models with an IMBH undergo core collapse at earlier times, the larger the IMBH mass the shallower, with an approximately constant central density at core collapse. The presence of an IMBH tends to lower the central velocity dispersion. These results hold independently of the mass function. For the models with Salpeter MF we observe that equipartition of kinetic energies is never achieved. Orbital anisotropy at large radii appears driven by energetic escapers on radial orbits. We measure the wander radius. Among the results we obtained, which mostly confirm or extend previously known trends that had been established over the range of parameters accessible to direct N-body simulations, we underline that the leptokurtic nature of the IMBH wander radius distribution might lead to IMBHs presenting as off-center more frequently than expected, with implications on observational IMBH detection.

astro-ph.GA↗

arXiv2022

Channel-driven Decentralized Bayesian Federated Learning for Trustworthy Decision Making in D2D Networks

Bayesian Federated Learning (FL) offers a principled framework to account for the uncertainty caused by limitations in the data available at the nodes implementing collaborative training. In Bayesian FL, nodes exchange information about local posterior distributions over the model parameters space. This paper focuses on Bayesian FL implemented in a device-to-device (D2D) network via Decentralized Stochastic Gradient Langevin Dynamics (DSGLD), a recently introduced gradient-based Markov Chain Monte Carlo (MCMC) method. Based on the observation that DSGLD applies random Gaussian perturbations of model parameters, we propose to leverage channel noise on the D2D links as a mechanism for MCMC sampling. The proposed approach is compared against a conventional implementation of frequentist FL based on compression and digital transmission, highlighting advantages and limitations.

eess.SP↗

arXiv2022

Symplectic coarse graining approach to the dynamics of spherical self-gravitating systems

We investigate the evolution of the phase-space distribution function around slightly perturbed stationary states and the process of violent relaxation in the context of the dissipationless collapse of an isolated spherical self-gravitating system. By means of the recently introduced symplectic coarse graining technique, we obtain an effective evolution equation that allows us to compute the scaling of the frequencies around a stationary state, as well as the damping times of Fourier modes of the distribution function, with the magnitude of the Fourier $k-$vectors themselves. We compare our analytical results with $N$-body simulations.

astro-ph.GA↗

arXiv2021

Opportunities of Federated Learning in Connected, Cooperative and Automated Industrial Systems

Next-generation autonomous and networked industrial systems (i.e., robots, vehicles, drones) have driven advances in ultra-reliable, low latency communications (URLLC) and computing. These networked multi-agent systems require fast, communication-efficient and distributed machine learning (ML) to provide mission critical control functionalities. Distributed ML techniques, including federated learning (FL), represent a mushrooming multidisciplinary research area weaving in sensing, communication and learning. FL enables continual model training in distributed wireless systems: rather than fusing raw data samples at a centralized server, FL leverages a cooperative fusion approach where networked agents, connected via URLLC, act as distributed learners that periodically exchange their locally trained model parameters. This article explores emerging opportunities of FL for the next-generation networked industrial systems. Open problems are discussed, focusing on cooperative driving in connected automated vehicles and collaborative robotics in smart manufacturing.

cs.LG↗

arXiv2018

The effects of shear and near tip deformations on interface fracture of symmetric sandwich beams

The effects of shear on energy release rate and mode mixity in a symmetric sandwich beam with isotropic layers and a debond crack at the face sheet/core interface are investigated through a semi-analytic approach based on two-dimensional elasticity and linear elastic fracture mechanics. Semi-analytic expressions are derived for the shear components of energy release rate and mode mixity phase angle which depend on four numerical coefficients derived through accurate finite element analyses. The expressions are combined with earlier results for three-layer configurations subjected to bending-moments and axial forces to obtain solutions for sandwich beams under general loading conditions and for an extensive range of geometrical and material properties. The results are applicable to laboratory specimens used for the characterization of the fracture properties of sandwich composites for civil, marine, energy and aeronautical applications, provided the lengths of the crack and the ligament ahead of the crack tip are above minimum lengths. The physical and mechanical significance of the terms of the energy release rate which depend on the shear forces are explained using structural mechanics concepts and introducing crack tip root rotations to account for the main effects of the near tip deformations.

cond-mat.soft↗