Lanzhou University of Technology Institutional Repository (LUT_IR)
OSLNet: Deep Small-Sample Classification with an Orthogonal Softmax Layer | |
Li, Xiaoxu1; Chang, Dongliang2; Ma, Zhanyu2; Tan, Zheng-Hua3; Xue, Jing-Hao1; Cao, Jie4; Yu, Jingyi1; Guo, Jun5 | |
2020 | |
会议录名称 | IEEE Transactions on Image Processing
![]() |
卷号 | 29 |
页码 | 6482-6495 |
出版者 | Institute of Electrical and Electronics Engineers Inc. |
摘要 | A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data. To mitigate overfitting in small-sample classification, learning more discriminative features from small-sample data is becoming a new trend. To this end, this paper aims to find a subspace of neural networks that can facilitate a large decision margin. Specifically, we propose the Orthogonal Softmax Layer (OSL), which makes the weight vectors in the classification layer remain orthogonal during both the training and test processes. The Rademacher complexity of a network using the OSL is only 1 K, where K is the number of classes, of that of a network using the fully connected classification layer, leading to a tighter generalization error bound. Experimental results demonstrate that the proposed OSL has better performance than the methods used for comparison on four small-sample benchmark datasets, as well as its applicability to large-sample datasets. Codes are available at: https://github.com/dongliangchang/OSLNet. © 1992-2012 IEEE. |
关键词 | Benchmarking Deep neural networks Large dataset TestingBenchmark datasets Discriminative features Function spaces Generalization error bounds Nonlinear layers Number of class Rademacher complexity Small sample datum |
DOI | 10.1109/TIP.2020.2990277 |
收录类别 | EI ; SCIE |
语种 | 英语 |
WOS研究方向 | Computer Science ; Engineering |
WOS类目 | Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic |
WOS记录号 | WOS:000545079400008 |
EI入藏号 | 20202808920107 |
EI主题词 | Multilayer neural networks |
ISSN | 10577149 |
来源库 | Compendex |
引用统计 | |
文献类型 | 会议论文 |
条目标识符 | https://ir.lut.edu.cn/handle/2XXMBERH/132673 |
专题 | 兰州理工大学 |
作者单位 | 1.School of Computer and Communication, Lanzhou University of Technology, Lanzhou; 730050, China; 2.Pattern Recognition and Intelligent System Laboratory, School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing; 730050, China; 3.Department of Electronic Systems, Aalborg University, Aalborg; 100876, Denmark; 4.Department of Statistical Science, University College London, London; 9220, United Kingdom; 5.School of Information Science and Technology, ShanghaiTech University, Shanghai; W1T 7PJ, China |
第一作者单位 | 兰州理工大学 |
推荐引用方式 GB/T 7714 | Li, Xiaoxu,Chang, Dongliang,Ma, Zhanyu,et al. OSLNet: Deep Small-Sample Classification with an Orthogonal Softmax Layer[C]:Institute of Electrical and Electronics Engineers Inc.,2020:6482-6495. |
条目包含的文件 | 条目无相关文件。 |
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。
修改评论