Actions
  • shareshare
  • link
  • cite
  • add
add
auto_awesome_motion View all 4 versions
Publication . Conference object . 2020

Fractional Step Discriminant Pruning: A Filter Pruning Framework for Deep Convolutional Neural Networks

Nikolaos Gkalelis; Vasileios Mezaris;
Open Access
Published: 06 Jul 2020
Publisher: IEEE
Abstract
In this paper, a novel pruning framework is introduced to compress noisy or less discriminant filters in small fractional steps, in deep convolutional networks. The proposed framework utilizes a class-separability criterion that can exploit effectively the labeling information in annotated training sets. Additionally, an asymptotic schedule for the pruning rate and scaling factor is adopted so that the selected filters’ weights collapse gradually to zero, providing improved robustness. Experimental results on the CIFAR-10, Google speech commands (GSC) and ImageNet32 (a downsampled version of ILSVRC-2012) show the efficacy of the proposed approach.
Subjects by Vocabulary

Microsoft Academic Graph classification: Pattern recognition Robustness (computer science) Schedule Discriminant Convolutional neural network Artificial intelligence business.industry business Computer science

Subjects

Deep convolutional neural networks, asymptotic filter pruning, class-separability criteria

moresidebar