Archivio Istituzionale della RicercaWe simulated tensor-train decomposed neural networks realized by Mach-Zehnder interferometer-based scalable photonic neuromorphic devices. The simulation results demonstrate that under practical hardware imprecisions, the TT-decomposed neural networks can achieve >90% test accuracy with 33.6× fewer MZIs than conventional photonic neural network implementations.
Analysis of the Hardware Imprecisions for Scalable and Compact Photonic Tensorized Neural Networks / On, M. B.; Lee, Y. -J.; Xiao, X.; Proietti, R.; Ben Yoo, S. J.. - ELETTRONICO. - (2021), pp. 1-4. (Intervento presentato al convegno 2021 European Conference on Optical Communication, ECOC 2021 tenutosi a Bordeaux, France nel 13-16 September 2021) [10.1109/ECOC52684.2021.9605948].
Analysis of the Hardware Imprecisions for Scalable and Compact Photonic Tensorized Neural Networks
Proietti R.;
2021
Abstract
We simulated tensor-train decomposed neural networks realized by Mach-Zehnder interferometer-based scalable photonic neuromorphic devices. The simulation results demonstrate that under practical hardware imprecisions, the TT-decomposed neural networks can achieve >90% test accuracy with 33.6× fewer MZIs than conventional photonic neural network implementations.
| File | Dimensione | Formato | |
|---|---|---|---|
|
Analysis_of_the_Hardware_Imprecisions_for_Scalable_and_Compact_Photonic_Tensorized_Neural_Networks.pdf
non disponibili
Tipologia:
2a Post-print versione editoriale / Version of Record
Licenza:
Non Pubblico - Accesso privato/ristretto
Dimensione
971.65 kB
Formato
Adobe PDF
|
971.65 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
|
ECOC2021_TPNN_simulation_v6.pdf
accesso aperto
Tipologia:
2. Post-print / Author's Accepted Manuscript
Licenza:
PUBBLICO - Tutti i diritti riservati
Dimensione
485.13 kB
Formato
Adobe PDF
|
485.13 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/11583/2973263