FROM FLEXIBLE AND BIOCOMPATIBLE DEVICES TO NEUROMORPHIC COMPUTING
- Anno accademico
- 2021/2022 Programmi anni precedenti
- Codice insegnamento
- PHD170 (AF:365494 AR:195058)
- Modalità
- In presenza
- Crediti formativi universitari
- 8
- Livello laurea
- Corso di Dottorato (D.M.45)
- Settore scientifico disciplinare
- ING-INF/01
- Periodo
- I Semestre
- Anno corso
- 1
- Sede
- VENEZIA
- Spazio Moodle
- Link allo spazio del corso
Inquadramento dell'insegnamento nel percorso del corso di studio
Prerequisiti
Contenuti
- micro and nanomaterials (CNTs, graphene and 2D materials, inorganic thin film) for flexible electronics;
- microfabrication of the devices: conventional microfabrication and additive manufacturing;
- challenge and limitations;
- examples of application: flexible displays, flexible sensors for brain monitoring and recording, epidermal circuits, degradable implants.
The second part of the course will focus on neuromorphic computing. Today, neural networks and machine learning algorithms are widely used to solve problems in data mining, vision recognition and language processing. These algorithms require high computational capabilities and are usually run on GPU (graphical processing unit). However, the computation on such machines is highly inefficient. New computer architectures that emulate the firing behaviour of the brain would be necessary to decrease power consumption and latency.
Here, some of the topics discussed in the course:
- computers Vs brain: Von Neuman (digital, clocked) and in-memory computing (analog, event-based)
- electrical model of the synapsis and its realization in CMOS technology;
- new devices and architecture for in-memory computing: resistive, phase change and ferroelectric memory;
- examples of application: computation, prosthetics, autonomous robots.
Testi di riferimento
- slides and additional readings will be provided during the course.