CASIA OpenIR  > 数字内容技术与服务研究中心  > 听觉模型与认知计算
Max Margin Cosine Loss for Speaker Identification on Short Utterances
Ji RF(吉瑞芳)1,2; Cao JH(曹俊华); Cai XY(蔡新元)1; Xu B(徐波)1
2018-11
Conference NameISCSLP 2018
Conference Date2018-11
Conference Place中国,台湾
Author of SourceISCSLP 2018
Publication Place中国,台湾
PublisherISCSLP 2018
Abstract

Speaker identification has made extraordinary progress owing to the advancement of deep neural networks. Speaker feature discrimination is a vital term in speaker recognition. However, the traditional softmax loss usually lacks the power of discrimination. To address this problem, this paper explores a novel loss function, namely max margin cosine loss (MMCL). To be specific, we realize the function by L2 normalizing both features and weight vectors in the softmax loss, together with a cosine margin term to maximize the decision margin in the angular space. In addition, max margin constraint, as one regularization term, is incorporated into the proposed loss function. Experimental results demonstrate the effectiveness of our proposed max margin cosine loss and superiority over pervious losses. For example, on 2s condition, MMCL reduces the equal error rate by 10.63% relatively compared to additive angular margin cosine loss (AMCL), while AMCL has already obtained 6.37% relative reduction than softmax loss.

Language英语
Document Type会议论文
Identifierhttp://ir.ia.ac.cn/handle/173211/23548
Collection数字内容技术与服务研究中心_听觉模型与认知计算
Affiliation1.中科院自动化研究所
2.中国科学院大学
First Author AffilicationInstitute of Automation, Chinese Academy of Sciences
Recommended Citation
GB/T 7714
Ji RF,Cao JH,Cai XY,et al. Max Margin Cosine Loss for Speaker Identification on Short Utterances[C]//ISCSLP 2018. 中国,台湾:ISCSLP 2018,2018.
Files in This Item: Download All
File Name/Size DocType Version Access License
ISCSLP.pdf(946KB)会议论文 开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Ji RF(吉瑞芳)]'s Articles
[Cao JH(曹俊华)]'s Articles
[Cai XY(蔡新元)]'s Articles
Baidu academic
Similar articles in Baidu academic
[Ji RF(吉瑞芳)]'s Articles
[Cao JH(曹俊华)]'s Articles
[Cai XY(蔡新元)]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Ji RF(吉瑞芳)]'s Articles
[Cao JH(曹俊华)]'s Articles
[Cai XY(蔡新元)]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: ISCSLP.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.