Davide Pilati

Dottorando in Ingegneria Elettrica, Elettronica E Delle Comunicazioni , 39o ciclo (2023-2026)
Dipartimento Scienza Applicata e Tecnologia (DISAT)

Docente esterno e/o collaboratore didattico
Dipartimento di Elettronica e Telecomunicazioni (DET)

Profilo

Dottorato di ricerca

Argomento di ricerca

Computing non convenzionale basato su reti neuromorfiche di nanofili

Tutori

Interessi di ricerca

Big Data, Machine Learning, Neural Networks and Data Science
Electronic devices: modeling and characterization
Micro- and nanotechnologies, devices, systems and applications

Biografia

Davide is a Ph.D. student in the field of electronics and nanotechnology, with a particular focus on nanocomputing and its applications in reservoir computing and neuromorphic nanowire networks. His academic journey began with an undergraduate degree in electronics engineering, where a course in digital electronics piqued his interest in neuromorphic computing. This initial intrigue evolved into a profound commitment to exploring the capabilities of nanowire networks in computational systems.

For his thesis, Davide developed and implemented a novel code to adapt a custom-designed electronic measurement board for use with nanowire networks. The heart of his thesis involved the multielectrode characterization of nanowire networks, revealing asymmetrical turn-on paths and tentative turn-ons. These findings provided new insights into the behavior of these networks and opened up avenues for further research.

One of Davide's most notable contributions to the field is the setup development for multielectrode characterization of these complex networks, which can be exploited for studying the emergent behavior of the system, as well as performing complex computational tasks.

As he looks towards his PhD, Davide is driven by ambitious goals. He aims to explore the potential of memristive nanowire networks further, with the hope of integrating this technology into systems for accelerated computation. Additionally, he is curious about the potential to exploit network nonlinearities to simplify computing architectures, potentially bypassing the need for a readout layer.

Didattica

Insegnamenti

Corso di laurea magistrale

Ricerca

Altre attività e progetti di ricerca

L'obiettivo principale della ricerca è l'implementazione sperimentale di paradigmi computazionali non convenzionali e di funzionalità neuromorfiche in dispositivi multiterminale basati su reti memristive di nanofili. Lo scopo è lo sviluppo di una architettura di computazione hardware capace di risolvere un ampio numero di task computazionali come riconoscimento di pattern e predizione di serie temporali con una riduzione in termini di costo di training e consumo energetico.

Pubblicazioni

Pubblicazioni più recenti Vedi tutte le pubblicazioni su Porto@Iris