Dynamical Conventional Neural Network Channel Pruning by Genetic Wavelet Channel Search for Image Classification
文献类型：期刊论文
| 作者 | Chen, Lin2 ; Gong, Saijun1; Shi, Xiaoyu2; Shang, Mingsheng2
|
| 刊名 | FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
 |
| 出版日期 | 2021-10-27 |
| 卷号 | 15页码:11 |
| 关键词 | neural network pruning neural architecture search wavelet features neural network compression image classification |
| DOI | 10.3389/fncom.2021.760554 |
| 通讯作者 | Shi, Xiaoyu(xiaoyushi@cigit.ac.cn) |
| 英文摘要 | Neural network pruning is critical to alleviating the high computational cost of deep neural networks on resource-limited devices. Conventional network pruning methods compress the network based on the hand-crafted rules with a pre-defined pruning ratio (PR), which fails to consider the variety of channels among different layers, thus, resulting in a sub-optimal pruned model. To alleviate this issue, this study proposes a genetic wavelet channel search (GWCS) based pruning framework, where the pruning process is modeled as a multi-stage genetic optimization procedure. Its main ideas are 2-fold: (1) it encodes all the channels of the pertained network and divide them into multiple searching spaces according to the different functional convolutional layers from concrete to abstract. (2) it develops a wavelet channel aggregation based fitness function to explore the most representative and discriminative channels at each layer and prune the network dynamically. In the experiments, the proposed GWCS is evaluated on CIFAR-10, CIFAR-100, and ImageNet datasets with two kinds of popular deep convolutional neural networks (CNNs) (ResNet and VGGNet). The results demonstrate that GNAS outperforms state-of-the-art pruning algorithms in both accuracy and compression rate. Notably, GNAS reduces more than 73.1% FLOPs by pruning ResNet-32 with even 0.79% accuracy improvement on CIFAR-100. |
| 资助项目 | National Nature Science Foundation of China[61902370] ; National Nature Science Foundation of China[61802360] ; Chongqing Research Program of Technology Innovation and Application[cstc2019jscx-zdztzxX0019] |
| WOS研究方向 | Mathematical & Computational Biology ; Neurosciences & Neurology |
| 语种 | 英语 |
| WOS记录号 | WOS:000717617100001 |
| 出版者 | FRONTIERS MEDIA SA |
| 源URL | [http://119.78.100.138/handle/2HOD01W0/14535]  |
| 专题 | 中国科学院重庆绿色智能技术研究院 |
| 通讯作者 | Shi, Xiaoyu |
| 作者单位 | 1.Tibet Univ, Sch Informat Sci & Technol, Lhasa, Peoples R China 2.Chinese Acad Sci, Chongqing Inst Green & Intelligent Technol, Chongqing, Peoples R China |
| 推荐引用方式 GB/T 7714 | Chen, Lin,Gong, Saijun,Shi, Xiaoyu,et al. Dynamical Conventional Neural Network Channel Pruning by Genetic Wavelet Channel Search for Image Classification[J]. FRONTIERS IN COMPUTATIONAL NEUROSCIENCE,2021,15:11. |
| APA | Chen, Lin,Gong, Saijun,Shi, Xiaoyu,&Shang, Mingsheng.(2021).Dynamical Conventional Neural Network Channel Pruning by Genetic Wavelet Channel Search for Image Classification.FRONTIERS IN COMPUTATIONAL NEUROSCIENCE,15,11. |
| MLA | Chen, Lin,et al."Dynamical Conventional Neural Network Channel Pruning by Genetic Wavelet Channel Search for Image Classification".FRONTIERS IN COMPUTATIONAL NEUROSCIENCE 15(2021):11. |
入库方式： OAI收割
来源：重庆绿色智能技术研究院
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。