KMS Of Academy of mathematics and systems sciences, CAS
Learning with smooth Hinge losses | |
Luo, JunRu1,2; Qiao, Hong3,4; Zhang, Bo5,6 | |
2021-11-06 | |
发表期刊 | NEUROCOMPUTING |
ISSN | 0925-2312 |
卷号 | 463页码:379-387 |
摘要 | Due to the non-smoothness of the Hinge loss in SVM, it is difficult to obtain a faster convergence rate with modern optimization algorithms. In this paper, we introduce two smooth Hinge losses psi(G)(alpha, sigma) and psi(M)(alpha; sigma) which are infinitely differentiable and converge to the Hinge loss uniformly in alpha as sigma tends to 0. By replacing the Hinge loss with these two smooth Hinge losses, we obtain two smooth support vector machines (SSVMs), respectively. Solving the SSVMs with the Trust Region Newton method (TRON) leads to two quadratically convergent algorithms. Experiments in text classification tasks show that the proposed SSVMs are effective in real-world applications. We also introduce a general smooth convex loss function to unify several commonly-used convex loss functions in machine learning. The general framework provides smooth approximation functions to non-smooth convex loss functions, which can be used to obtain smooth models that can be solved with faster convergent optimization algorithms. (C) 2021 Elsevier B.V. All rights reserved. |
关键词 | Smooth Hinge loss Convex surrogate loss Support vector machine Trust region Newton method |
DOI | 10.1016/j.neucom.2021.08.060 |
收录类别 | SCI |
语种 | 英语 |
WOS研究方向 | Computer Science |
WOS类目 | Computer Science, Artificial Intelligence |
WOS记录号 | WOS:000708073900014 |
出版者 | ELSEVIER |
引用统计 | |
文献类型 | 期刊论文 |
条目标识符 | http://ir.amss.ac.cn/handle/2S8OKBNM/59420 |
专题 | 应用数学研究所 |
通讯作者 | Zhang, Bo |
作者单位 | 1.Changzhou Univ, Sch Comp Sci & Artificial Intelligence, Changzhou 213164, Jiangsu, Peoples R China 2.Changzhou Univ, Aliyun Sch Big Data, Changzhou 213164, Jiangsu, Peoples R China 3.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China 4.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 5.Chinese Acad Sci, Acad Math & Syst Sci, Beijing 100190, Peoples R China 6.Univ Chinese Acad Sci, Sch Math Sci, Beijing 100049, Peoples R China |
推荐引用方式 GB/T 7714 | Luo, JunRu,Qiao, Hong,Zhang, Bo. Learning with smooth Hinge losses[J]. NEUROCOMPUTING,2021,463:379-387. |
APA | Luo, JunRu,Qiao, Hong,&Zhang, Bo.(2021).Learning with smooth Hinge losses.NEUROCOMPUTING,463,379-387. |
MLA | Luo, JunRu,et al."Learning with smooth Hinge losses".NEUROCOMPUTING 463(2021):379-387. |
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