204 research outputs found
Informe científico de investigador: Paolini, Eduardo Emilio (2012-2013)
No full textDurante este período el trabajo de investigación se orientó en dos líneas principales:
• reducción de la distorsión en amplificadores conmutados, y
• reducción de ruido y distorsión en sistemas de lectura de CCDs (Charge Coupled Devices).
utilizando técnicas de procesamiento digital de señales.
En la primera línea se buscó reducir el contenido armónico en una banda de frecuencias de interés cuando se trabaja con señales tipo binarias (“1”/”0”), que son adecuadas para desarrollar equipos electrónicos con una alta eficiencia energética. Estos resultados son aplicables a convertidores de potencia AC/AC o DC/AC, amplificadores de potencia, etc.
Los resultados alcanzados en este período permitieron reducir la distorsión de algoritmos de modulación por ancho de pulsos (MAP o PWM por sus siglas en inglés) desarrollando dos tipos de técnicas diferentes: una denominada “modulación click” (trabajo 3) y otra conocida como “MAP con distorsión cero en banda base” (trabajo 4). En todos los casos, se busca diseñar algoritmos de tiempo discreto, que puedan ejecutarse en procesadores digitales o FPGAs, y no en el desarrollo de electrónica analógica. El enfoque de cada una de estas técnicas es distinto, pero el resultado es similar. En el primer caso (modulación click) se plantea el problema en el dominio frecuencia (buscar que el espectro de la señal modulada sea idéntico al de la señal moduladora en la banda de interés), y en el segundo se parte en el dominio temporal (hacer que las muestras de la señal modulada sean idénticas a la de la señal moduladora), resultando en algoritmos distintos que satisfacen los requisitos de diseño. Durante este período las investigaciones estuvieron orientadas a obtener versiones prácticas e implementables de esos algoritmos. Tales aspiraciones fueron cumplidas satisfactoriamente, y sus resultados se publicaron en los trabajos 3 y 4, y han dado lugar a otra serie de trabajos actualmente en prensa (trabajos 9, 10, 11, 12 y 15).
Con respecto a la segunda línea de investigación, originada por una colaboración con investigadores del Fermi National Accelerator Laboratory, Batavia, Illinois, EE UU., el propósito es utilizar sensores tipo CCD (Charge Coupled Devices), semejantes a los que se utilizan en cámaras digitales, para detectar partículas elementales. Estos sensores se están utilizando para dos tipos de experimentos, que son detectar la presencia de materia oscura en el universo (experimento DAMIC) y la detección de neutrinos de bajas energías (experimento CONNIE), que permitiría en principio monitorear el estado de reactores atómicos, específicamente el de la central nuclear Almirante Alvaro Alberto en Angra do Reis, Brasil. Los CCD para aplicaciones científicas tienen alta eficiencia de detección, buena resolución espacial, bajo ruido, y son más gruesos para facilitar la interacción con partículas. De todos modos, el nivel de interacción es tan bajo que es fundamental aplicar técnicas que permitan reducir el ruido de lectura de tales dispositivos La colaboración de nuestro grupo está vinculada al procesamiento de las señales provenientes del CCD, incluyendo la generación de relojes, muestreo, análisis e interpretación de resultados. En los trabajos 1, 6 y 7 se estudian algunas técnicas de procesamiento que permiten alcanzar niveles de ruido compatibles con este tipo de aplicaciones, y en el trabajo 16 se estudia la factibilidad del empleo de este tipo de sensores para la detección no ambigua de neutrinos de baja energía
Informe científico de investigador: Paolini, Eduardo Emilio (2016-2017)
No full textEl área de trabajo es el análisis, modelado y control de sistemas electrónicos, y en particular la aplicación de técnicas de procesamiento de señales para reducir el ruido y/o eliminar la distorsión en esta clase de sistemas. Estas líneas de trabajo se aplican en la síntesis de señales analógicas utilizando señales binarias, en el desarrollo de técnicas de modulación de ancho de pulso sin distorsión en banda base para amplificadores de potencia para audio y comunicaciones, y en la lectura de CCDs científicos para la detección de neutrinos generados por reactores nucleares, como parte de una colaboración con el Fermi National Laboratory (Chicago, EE.UU.)
Informe científico de investigador: Paolini, Eduardo Emilio (2014-2015)
No full textEl área de trabajo es el análisis, modelado y control de sistemas electrónicos, y en particular la aplicación de técnicas de procesamiento de señales para reducir el ruido y/o eliminar la distorsión en esta clase de sistemas. Estas líneas de trabajo se aplican en la síntesis de señales analógicas utilizando señales binarias, en el desarrollo de técnicas de modulación de ancho de pulso sin distorsión en banda base para amplificadores de potencia para audio y comunicaciones, y en la lectura de CCDs científicos para la detección de neutrinos generados por reactores nucleares, como parte de una colaboración con el Fermi National Laboratory (Chicago, EE.UU.)
The diagnostic technique of False Colour and the application of Artificial Intelligence for its interpretation
No full textLa tecnica diagnostica del Falso Colore, sia in radiazione infrarossa che nel campo dell’ultravioletto, permette di individuare con precisione lo stato di conservazione di un dipinto, ma la potenziale riconoscibilità dei pigmenti che compongono la tavolozza cromatica attraverso questo tipo di esame è estremamente difficile, sia per la difficoltà di distinguere le miscele di più dipinti, sia per l’estrema soggettività della lettura. Il saggio presenta un’inedita metodologia, progettata dall’autore del contributo, legata al Falso Colore analizzato attraverso un sistema integrato di Intelligenza Artificiale.The diagnostic technique of False Colour, both in infrared and ultraviolet radiation, makes it possible to accurately identify the state of conservation of a painting, but the potential recognisability of the pigments that make up the colour palette through this type of examination is extremely difficult, due both to the difficulty of distinguishing mixtures of several paintings and to the extreme subjectivity of the reading. The paper presents a novel methodology, designed by the author of the contribution, related to False Colour analysed through an integrated Artificial Intelligence system
Artificial intelligence applied to the analysis of paintings: between false myths and new perspectives
No full textL’idea di utilizzare l’Intelligenza Artificiale per la diagnostica artistica non è una novità: finora, però, la sua applicazione è stata piegata, quasi esclusivamente, alla utopistica certificazione di autenticità delle opere d’arte, pretendendo di sostituire la sensibilità umana nel riconoscimento dei tratti stilistici con l’automazione. Questo approccio metodologico è errato poiché non è in grado di rispondere alla variabile umana con cui ogni opera d’arte è stata realizzata. L’utilizzo dell’Intelligenza Artificiale, però, è funzionale all’indagine analitica di fattori diagnostici che, se letti da occhio umano, possono dare adito a interpretazioni non necessarie e fuorvianti: in particolare, una lettura sistemica e automatica dei risultati in Falso Colore, può restituire una lettura veritiera e completa dei pigmenti che compongono la superficie di un dipinto. Questa metodologia di indagine, ideata dall’autore dell’articolo e basata su un sistema di tre metodi algoritmici differenti, è stata testata su circa trecento dipinti, la cui tavolozza cromatica è stata controllata anche attraverso le tradizionali tecniche spettrografiche.The idea of using Artificial Intelligence for art diagnostics is nothing new. Until now, however, its application has been bent, almost exclusively, to the utopian certification of the authenticity of works of art, claiming to replace human sensitivity in the recognition of stylistic traits with automation. This methodological approach is flawed as it is unable to respond to the human variable with which each work of art was created. The use of Artificial Intelligence, however, is functional to the analytical investigation of diagnostic factors that, if read by the human eye, can give rise to unnecessary and misleading interpretations: in particular, a systemic and automatic reading of the results in False Colour, can return a true and complete reading of the pigments that make up the surface of a painting. This investigation methodology, devised by the author of the article and based on a system of three different algorithmic methods, was tested on approximately three hundred paintings, whose colour palette was also checked using traditional spectrographic techniques
Filippo Neri e Felice da Cantalice figli della Vergine. L'umorismo al servizio della carità per la devozione mariana nella Roma di fine '500
No full textNella Roma post-conciliare, l’esigenza di legare la popolazione alla dottrina di Santa Romana Chiesa divenne sempre più urgente: era necessario arginare le conversioni verso la cristianità riformata attraverso esempi concreti e interventi materiali a favore dei bisognosi, al fine di dimostrare la prodigalità della chiesa cattolica. In questo contesto, si devono inquadrare le attività caritatevoli, ma anche didattiche e divulgative, di Filippo Neri e Felice da Cantalice, di formazione domenicana e fondatore della Congregazione dell’Oratorio il primo, frate cappuccino il secondo. Il rivoluzionario metodo educativo dei due religiosi, basato non sulla coercizione e sulle punizioni, come tipico dell’epoca, ma sulla gioia, il riso e la condivisione, mirava all’esaltazione della Vergine in quanto Sedes Sapientiae, e quindi dispensatrice di Carità. Sotto questo punto di vista, il saggio analizza le iconografie che raffigurano Filippo Neri e Felice da Cantalice nelle intrinseche analogie e differenze, specificandone i significati iconologici nel contesto della diffusione del culto mariano cattolico. In particolare, l’esperienza mistica che entrambi i religiosi vissero nei confronti della Vergine con il Bambino, e che divenne il soggetto principale della loro rappresentazione, trova una matrice comune nell’antica icona bizantina della Panaghia Nikopoia.In post-conciliar Rome, the need to bind the population to the doctrine of the Holy Roman Church became more and more urgent: it was necessary to stem conversions towards reformed Christianity through concrete examples and material interventions in favour of the needy, in order to demonstrate the prodigality of the Catholic Church. In this context, one must place the charitable activities, but also the educational and popular activities, of Philip Neri and Felix of Cantalice, of Dominican formation and founder of the Congregation of the Oratory the former, Capuchin friar the latter. The revolutionary educational method of the two religious, based not on coercion and punishment, as was typical of the time, but on joy, laughter and sharing, aimed at the exaltation of the Virgin as Sedes Sapientiae, and therefore dispenser of Charity. From this point of view, the essay analyses the iconographies depicting Philip Neri and Felix of Cantalice in their intrinsic similarities and differences, specifying their iconological meanings in the context of the spread of the Catholic Marian cult. In particular, the mystical experience that both religious men had with the Virgin and Child, and which became the main subject of their representation, finds a common matrix in the ancient Byzantine icon of the Panaghia Nikopoia
Enhancing Neuromorphic Photonic Hardware Performance through Neural Architecture Search
No full textCurrent neuromorphic photonic neural networks cannot achieve the performance of electronic neural networks due to the presence of physical constraints, such as noise and distortions, affecting the implementations based on analog photonic hardware. This paper proposes the exploitation of Neural Architecture Search (NAS) tailored for Photonic-Aware Neural Networks (PANNs), a class of neural networks amenable to a neuromorphic hardware implementation. In this way we are able to obtain PANN architectures that not only meet these photonic constraints but also to outperform existing photonic models. Experimental results on the CIFAR-10 dataset demonstrate that exploiting NAS while addressing photonic constraints can signifi cantly improve the performance of PANNs, obtaining results that are comparable with state-of-the-art electronic networks. Indeed, the best-performing configuration achieved an accuracy of 95%, a performance similar to electronic counterparts while complying with photonic constraints
Real-time clustering based on deep embeddings for threat detection in 6G networks
No full textTrials and deployments of sixth Generation (6G) wireless networks, delivering extreme capacity, reliability, and effi ciency, are expected as early as 2030. Attempts from both industry and academia are trying to defi ne the next generation network infrastructure. 6G will set in motion the fourth industrial revolution, delivering global, integrated intelligence. In this scenario, Artifi cial Intelligence (AI)-assisted architecture for 6G networks will realize knowledge discovery, automatic network adjustment and intelligent service provisioning. The long-term vision for implementing 6G security is to implement an autonomous, self-learning AI-assisted architecture that can perform threat mitigation without disrupting the normal use, enabling the resilience and reliability of the network and fulfi lling security automation. This work proposes
a fi rst implementation of a proactive threat discovery service to be deployed at 6G base stations, paving the way for collective network intelligence in the context of cybersecurity mechanisms. Specifi cally, a fully unsupervised Deep Learning (DL) model is presented, able to recognize both Denial of Service (DoS) Hulk and DoS Goldeneye, with 97 . 0% and 92 . 2% mean F1-score respectively
Real-Time Network Packet Classification Exploiting Computer Vision Architectures
No full textForthcoming 6G/NextG networks highlight the need for advanced Artificial Intelligence (AI)-based security mechanisms to identify malicious activities and adapt to emerging threats. In this context, the integration of computer vision techniques into the cybersecurity field is promising due to their potential for sophisticated pattern recognition. In this paper we introduce a computationally efficient classification scheme acting directly on the raw packets collected at base stations and enforcing real-time conversion of packets into images. The innovative points of the proposed solution are the lightweight implementation, aligning well with the demands of future 6G networks, and the operation at network edge, enabling early threat identification as close as possible to the packet origin. We investigate the performance of this approach both in terms of F1-score and prediction time using state-of-the-art computer vision architectures and a customized Convolutional Neural Network (CNN) in an intrusion detection problem using a 5G dataset. Experimental results show the superiority of the CNN architecture over complex models. Across multiple packet window sizes (i.e., 10, 50, 100 packets), the CNN consistently outperforms the other state-of-the-art computer vision models, achieving very high F1-scores (namely, 0.99593, 0.99860, 0.99895). A scalability analysis highlights a trade-off between CNN scalability and performance, where larger values lead to increased prediction time. On the other hand, the other computer vision models exhibit better scalability, enabling an optimal model selection without trade-offs
Elimination of dead-time distortion using time-feedback
No full textA fully-digital algorithm to shape the spectrum of dead-time distortion in power inverters is presented. Dead-time is required to avoid short circuits of the power source by the legs of a power inverter due to the finite turn-on and turn-off times of the switches. Dead-time modifies the pulse widths of the PWM signal causing high harmonic distortion. The proposed approach is based on measuring the pulse widths at the output of the power stage, and does not require to measure the sign of the current, which is usually necessary for dead-time compensation algorithms. Furthermore, the quantization noise produced by digital PWM is also reduced, and it has a minimal computational cost.Fil: Chierchie, Fernando. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca. Instituto de Investigaciones en Ingeniería Eléctrica "Alfredo Desages". Universidad Nacional del Sur. Departamento de Ingeniería Eléctrica y de Computadoras. Instituto de Investigaciones en Ingeniería Eléctrica "Alfredo Desages"; ArgentinaFil: Paolini, Eduardo Emilio. Universidad Nacional del Sur. Departamento de Ingeniería Eléctrica y de Computadoras. Instituto "alfredo Desages"; ArgentinaXVII Workshop on Information Processing and ControlMar del PlataArgentinaUniversidad Nacional de Mar del Plat
- …
