Timezone: »

 
Poster
PerforatedCNNs: Acceleration through Elimination of Redundant Convolutions
Mikhail Figurnov · Aizhan Ibraimova · Dmitry Vetrov · Pushmeet Kohli

Wed Dec 07 09:00 AM -- 12:30 PM (PST) @ Area 5+6+7+8 #42
in Wednesday Posters »

We propose a novel approach to reduce the computational cost of evaluation of convolutional neural networks, a factor that has hindered their deployment in low-power devices such as mobile phones. Inspired by the loop perforation technique from source code optimization, we speed up the bottleneck convolutional layers by skipping their evaluation in some of the spatial positions. We propose and analyze several strategies of choosing these positions. We demonstrate that perforation can accelerate modern convolutional networks such as AlexNet and VGG-16 by a factor of 2x - 4x. Additionally, we show that perforation is complementary to the recently proposed acceleration method of Zhang et al.

Keywords: Deep Learning or Neural Networks · (Application) Computer Vision
[ PDF] 

Author Information

Mikhail Figurnov (Skolkovo Inst. of Sc and Tech)
Aizhan Ibraimova (Skolkovo Institute of Science and Technology)
Dmitry Vetrov (Higher School of Economics, AI Research Institute)
Pushmeet Kohli (Microsoft Research)

More from the Same Authors