An efficient binary search based neuron pruning method for ConvNet condensation

Zhang, B, Qin, K and Chan, J 2017, 'An efficient binary search based neuron pruning method for ConvNet condensation', in D. Liu, S. Xie, Y. Li, D. Zhao and E.-S. M. El-Alfy (ed.) Proceedings of the 24th International Conference on Neural Information Processing, Guangzhou, China, 14-18 November 2017, pp. 189-197.


Document type: Conference Paper
Collection: Conference Papers

Title An efficient binary search based neuron pruning method for ConvNet condensation
Author(s) Zhang, B
Qin, K
Chan, J
Year 2017
Conference name ICONIP 2017: International Conference on Neural Information Processing
Conference location Guangzhou, China
Conference dates 14-18 November 2017
Proceedings title Proceedings of the 24th International Conference on Neural Information Processing
Editor(s) D. Liu, S. Xie, Y. Li, D. Zhao and E.-S. M. El-Alfy
Publisher Springer International Publishing
Place of publication United States
Start page 189
End page 197
Total pages 9
Abstract Convolutional neural networks (CNNs) have been widely applied in the field of computer vision. Nowadays, the architecture of CNNs is becoming more and more complex, involving more layers and more neurons per layer. The augmented depth and width of CNNs will lead to greatly increased computational and memory costs, which may limit CNNs practical utility. However, as demonstrated in previous research, CNNs of complex architecture may contain considerable redundancy in terms of hidden neurons. In this work, we propose a magnitude based binary neuron pruning method which can selectively prune neurons to shrink the network size while keeping the performance of the original model without pruning. Compared to some existing neuron pruning methods, the proposed method can achieve higher compression rate while automatically determining the number of neurons to be pruned per hidden layer in an efficient way.
Subjects Neural, Evolutionary and Fuzzy Computation
Pattern Recognition and Data Mining
Keyword(s) deep network
Copyright notice © Springer International Publishing AG 2017
ISBN 9783319700878
Versions
Version Filter Type
Access Statistics: 44 Abstract Views  -  Detailed Statistics
Created: Mon, 04 Dec 2017, 13:07:00 EST by Catalyst Administrator
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us