Reduced-Order Modeling of Deep Neural Networks

We introduce a new method for speeding up the inference of deep neural networks. It is somewhat inspired by the reduced-order modeling techniques for dynamical systems. The cornerstone of the proposed method is the maximum volume algorithm. We demonstrate efficiency on neural networks pre-trained on...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	Computational mathematics and mathematical physics 2021-05, Vol.61 (5), p.774-785
Hauptverfasser:	Gusak, J., Daulbaev, T., Ponomarev, E., Cichocki, A., Oseledets, I.
Format:	Artikel
Sprache:	eng
Schlagworte:	Algorithms Artificial neural networks Computational Mathematics and Numerical Analysis General Numerical Methods Mathematics Mathematics and Statistics Model reduction Modelling Neural networks Reduced order models
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!