Source author record

Alessandro Beghi

Alessandro Beghi appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

physics.optics Artificial Intelligence Machine Learning physics.ins-det

Catalog footprint

What is connected

4works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

A Multi-label Continual Learning Framework to Scale Deep Learning Approaches for Packaging Equipment Monitoring

Continual Learning aims to learn from a stream of tasks, being able to remember at the same time both new and old tasks. While many approaches were proposed for single-class classification, multi-label classification in the continual scenario remains a challenging problem. For the first time, we study multi-label classification in the Domain Incremental Learning scenario. Moreover, we propose an efficient approach that has a logarithmic complexity with regard to the number of tasks, and can be applied also in the Class Incremental Learning scenario. We validate our approach on a real-world multi-label Alarm Forecasting problem from the packaging industry. For the sake of reproducibility, the dataset and the code used for the experiments are publicly available.

preprint2016arXiv

Computing the laser beam path in optical cavities: a geometric Newton's method based approach

In the last decade, increasing attention has been drawn to high precision optical experiments, which push resolution and accuracy of the measured quantities beyond their current limits. This challenge requires to place optical elements (e.g. mirrors, lenses, etc.) and to steer light beams with sub-nanometer precision. Existing methods for beam direction computing in resonators, e.g. iterative ray tracing or generalized ray transfer matrices, are either computationally expensive or rely on overparametrized models of optical elements. By exploiting Fermat's principle, we develop a novel method to compute the steady-state beam configurations in resonant optical cavities formed by spherical mirrors, as a function of mirror positions and curvature radii. The proposed procedure is based on the geometric Newton method on matrix manifold, a tool with second order convergence rate that relies on a second order model of the cavity optical length. As we avoid coordinates to parametrize the beam position on mirror surfaces, the computation of the second order model does not involve the second derivatives of the parametrization. With the help of numerical tests, we show that the convergence properties of our procedure hold for non-planar polygonal cavities, and we assess the effectiveness of the geometric Newton method in determining their configurations with high degree of accuracy and negligible computational effort.

preprint2013arXiv

Controlling the nonlinear intracavity dynamics of large He-Ne laser gyroscopes

A model based on Lamb's theory of gas lasers is applied to a He-Ne ring laser gyroscope in order to estimate and remove the laser dynamics contribution from the rotation measurements. The intensities of the counter-propagating laser beams exiting one cavity mirror are continuously observed together with a monitor of the laser population inversion. These observables, once properly calibrated with a dedicated procedure, allow us to estimate cold cavity and active medium parameters driving the main part of the nonlinearities of the system. The parameters identification and noise subtraction procedure has been verified by means of a Monte Carlo study of the system, and experimentally tested on the G-Pisa ring laser oriented with the normal to the ring plane almost parallel to the Earth rotation axis. In this configuration the Earth rotation-rate provides the maximum Sagnac effect while the contribution of the orientation error is reduced at minimum. After the subtraction of laser dynamics by a Kalman filter, the relative systematic errors of G-PISA reduce from 50 to 5 part in 10^3 and can be attributed to the residual uncertainties on geometrical scale factor and orientation of the ring.

preprint2012arXiv

Compensation of the laser parameters fluctuations in large ring laser gyros: a Kalman filter approach

He-Ne ring laser gyroscopes are, at present, the most precise devices for absolute angular velocity measurements. Limitations to their performance come from the non--linear dynamics of the laser. Following the Lamb semi-classical theory, we find a set of critical parameters affecting the time stability of the system. We propose a method for estimating the long term drift of the laser parameters and for filtering out the laser dynamics effects from the rotation measurement. The parameter estimation procedure, based on the perturbative solutions of the laser dynamics, allow us to apply Kalman Filter theory for the estimation of the angular velocity. Results of a comprehensive Monte Carlo simulation and results of a preliminary analysis on experimental data from the ring laser prototype G-Pisa are shown and discussed.