The thesis studies resistive switching and conductive bridge devices for memory and neuromorphic applications. It focuses on physical mechanisms of resistive switching and reliability issues, in particular device variability, random telegraph noise and noise issues at both single cell and array levels, cycling endurance and degradation. All these aspects are physically modeled, allowing for a deeper understanding of the device behavior. Finally, the thesis provides novel computing approaches by employing the resistive switching device as a synapse for hardware neuromorphic networks for pattern learning and recognition.
Il lavoro di tesi studia le memorie a switching resistivo e a ponte conduttivo per applicazioni di memoria e in reti neuromorfiche. Ci si concentra sui meccanismi fisici di switching resistivo e sui problemi di affidabilità, in particolare variabilità del dispositivo, rumore telegrafico e altri problemi di rumore, sia a livello di singola cella che di array, ciclatura e degrado. Tutti questi aspetti sono modellizzati tramite un approccio fisico, consentendo una comprensione più approfondita del funzionamento del dispositivo. Infine, la tesi propone nuovi approcci di calcolo impiegando le memorie resistive come sinapsi per reti neuromorfiche hardware per apprendimento e riconoscimento di pattern.
Modeling of reliability and neuromorphic application of resistive switching devices
AMBROGIO, STEFANO
Abstract
The thesis studies resistive switching and conductive bridge devices for memory and neuromorphic applications. It focuses on physical mechanisms of resistive switching and reliability issues, in particular device variability, random telegraph noise and noise issues at both single cell and array levels, cycling endurance and degradation. All these aspects are physically modeled, allowing for a deeper understanding of the device behavior. Finally, the thesis provides novel computing approaches by employing the resistive switching device as a synapse for hardware neuromorphic networks for pattern learning and recognition.File | Dimensione | Formato | |
---|---|---|---|
2015_PhD_Thesis_Stefano_Ambrogio.pdf
non accessibile
Descrizione: Tesi
Dimensione
63.9 MB
Formato
Adobe PDF
|
63.9 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/117782