arXiv subjects

B. Paroli

Publications and source records attributed to B. Paroli.

4 recordsLinked to original sources

arXiv2025

The receptron is a nonlinear threshold logic gate with intrinsic multi-dimensional selective capabilities for analog inputs

Threshold logic gates (TLGs) have been proposed as artificial counterparts of biological neurons with classification capabilities based on a linear predictor function combining a set of weights with the feature vector. The linearity of TLGs limits their classification capabilities requiring the use of networks for the accomplishment of complex tasks. A generalization of the TLG model called receptron, characterized by input-dependent weight functions allows for a significant enhancement of classification performances even with the use of a single unit. Here we formally demonstrate that a receptron, characterized by nonlinear input-dependent weight functions, exhibit intrinsic selective activation properties for analog inputs, when the input vector is within cubic domains in a 3D space. The proposed model can be extended to the n-dimensional case for multidimensional applications. Our results suggest that receptron-based networks can represent a new class of devices capable to manage a large number of analog inputs, for edge applications requiring high selectivity and classification capabilities without the burden of complex training.

cs.ET↗

arXiv2022

Solving classification tasks by a receptron based on nonlinear optical speckle fields

Among several approaches to tackle the problem of energy consumption in modern computing systems, two solutions are currently investigated: one consists of artificial neural networks (ANNs) based on photonic technologies, the other is a different paradigm compared to ANNs and it is based on random networks of nonlinear nanoscale junctions resulting from the assembling of nanoparticles or nanowires as substrates for neuromorphic computing. These networks show the presence of emergent complexity and collective phenomena in analogy with biological neural networks characterized by self-organization, redundancy, non-linearity. Starting from this background, we propose and formalize a generalization of the perceptron model to describe a classification device based on a network of interacting units where the input weights are nonlinearly dependent. We show that this model, called "receptron", provides substantial advantages compared to the perceptron as, for example, the solution of non-linearly separable Boolean functions with a single device. The receptron model is used as a starting point for the implementation of an all-optical device that exploits the non-linearity of optical speckle fields produced by a solid scatterer. By encoding these speckle fields we generated a large variety of target Boolean functions without the need for time-consuming machine learning algorithms. We demonstrate that by properly setting the model parameters, different classes of functions with different multiplicity can be solved efficiently. The optical implementation of the receptron scheme opens the way for the fabrication of a completely new class of optical devices for neuromorphic data processing based on a very simple hardware.

cond-mat.dis-nn↗

arXiv2012

Thomson backscattering diagnostics of nanosecond electron bunches in high space charge regime

The intra-beam repulsions play a significant role in determining the performances of free-electron devices when an high brilliance of the beam is required. The transversal and longitudinal spread of the beam, its energy and density are fundamental parameters in any beam experiment and different beam diagnostics are available to measure such parameters. A diagnostic method based on the Thomson backscattering of a laser beam impinging on the particle beam is proposed in this work for the study of nanosecond electron bunches in high space charge regime. This diagnostics, aimed to the measurement of density, energy and energy spread, was set-up in a Malmberg-Penning trap (generally used for the electron/ion confinment) in two different configurations designed to optimize sensitivity, spatial resolution and electron-beam coincidence in space and time. To this purpose an electron bunch (pulse time <4ns), produced by a photocathode source, was preliminary characterized with different electrostatic diagnostics and used to test the diagnostics systems. The solutions are detailed, which were devised for both the laser and bunch injection in the vacuum chamber, space and time coincidence of electron and laser pulses, photon detection, optimization of the geometry in the laser-beam interaction. The results are then summarized with an estimate of the minimum sensitivity of the set-up.

physics.plasm-ph↗

arXiv2012

A coherent polarimeter array for the Large Scale Polarization Explorer balloon experiment

We discuss the design and expected performance of STRIP (STRatospheric Italian Polarimeter), an array of coherent receivers designed to fly on board the LSPE (Large Scale Polarization Explorer) balloon experiment. The STRIP focal plane array comprises 49 elements in Q band and 7 elements in W-band using cryogenic HEMT low noise amplifiers and high performance waveguide components. In operation, the array will be cooled to 20 K and placed in the focal plane of a $\sim 0.6$ meter telescope providing an angular resolution of $\sim1.5$ degrees. The LSPE experiment aims at large scale, high sensitivity measurements of CMB polarization, with multi-frequency deep measurements to optimize component separation. The STRIP Q-band channel is crucial to accurately measure and remove the synchrotron polarized component, while the W-band channel, together with a bolometric channel at the same frequency, provides a crucial cross-check for systematic effects.

astro-ph.IM↗