Impact of Disentanglement on Pruning Neural Networks

[en] Efficient model compression techniques are required to deploy deep neural networks (DNNs) on edge devices for task specific objectives. A variational autoencoder (VAE) framework is combined with a pruning criterion to investigate the impact of having the network learn disentangled representations on the pruning process for the classification task.

Research center :

Interdisciplinary Centre for Security, Reliability and Trust (SnT) > CVI² - Computer Vision Imaging & Machine Intelligence

Disciplines :

Physical, chemical, mathematical & earth Sciences: Multidisciplinary, general & others
Electrical & electronics engineering
Engineering, computing & technology: Multidisciplinary, general & others

Author, co-author :

SHNEIDER, Carl ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

ROSTAMI ABENDANSARI, Peyman ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

KACEM, Anis ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

SINHA, Nilotpal ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

SHABAYEK, Abd El Rahman ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

AOUADA, Djamila ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2

External co-authors :

Language :

English

Title :

Impact of Disentanglement on Pruning Neural Networks

Publication date :

20 June 2023

Number of pages :

Event name :

International Symposium on Computational Sensing (ISCS23)

Event organizer :

Thomas Feuillen, Amirafshar Moshtaghpour

Event place :

Luxembourg, Luxembourg

Event date :

12-06-2023 to 14-06-2023

Audience :

International

Focus Area :

Computational Sciences
Security, Reliability and Trust

Additional URL :

https://zenodo.org/record/8060030

FnR Project :

FNR15965298 - Enabling Learning And Inferring Compact Deep Neural Network Topologies On Edge Devices, 2021 (01/09/2022-31/08/2025) - Djamila Aouada

Name of the research project :

Enabling Learning And Inferring Compact Deep Neural Network Topologies On Edge Devices (ELITE)

Funders :

FNR - Fonds National de la Recherche [LU]

Available on ORBilu :

since 02 October 2023

Statistics

Number of views

34 (3 by Unilu)

Number of downloads

12 (0 by Unilu)

More statistics