Geometry Preserving Multi-task Metric Learning
Yang, Peipei; Huang, Kaizhu; Liu, Cheng-Lin
2012-09
会议名称European Conference on Machine Learning
会议录名称Machine Learning and Knowledge Discovery in Databases
会议日期2012-9-23
会议地点Bristol, UK
摘要Multi-task learning has been widely studied in machine learning due to its capability to improve the performance of multiple related learning problems. However, few researchers have applied it on the important metric learning problem. In this paper, we propose to couple multiple related metric learning tasks with von Neumann divergence. On one hand, the novel regularized approach extends previous methods from the vector regularization to a general matrix regularization framework; on the other hand and more importantly, by exploiting von Neumann divergence as the regularizer, the new multi-task metric learning has the capability to well preserve the data geometry. This leads to more appropriate propagation of side-information among tasks and provides potential for further improving the performance. We propose the concept of geometry preserving probability (PG) and show that our framework leads to a larger PG in theory. In addition, our formulation proves to be jointly convex and the global optimal solution can be guaranteed. A series of experiments across very different disciplines verify that our proposed algorithm can consistently outperform the current methods.
关键词Multi-task Learning Metric Learning Geometry Preserving
DOI10.1007/978-3-642-33460-3_47
引用统计
文献类型会议论文
条目标识符http://ir.ia.ac.cn/handle/173211/12500
专题模式识别国家重点实验室_模式分析与学习
通讯作者Huang, Kaizhu
作者单位National Laboratory of Pattern Recognition
推荐引用方式
GB/T 7714
Yang, Peipei,Huang, Kaizhu,Liu, Cheng-Lin. Geometry Preserving Multi-task Metric Learning[C],2012.
条目包含的文件 下载所有文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
chp%3A10.1007%2F978-(470KB)会议论文 开放获取CC BY-NC-SA浏览 下载
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[Yang, Peipei]的文章
[Huang, Kaizhu]的文章
[Liu, Cheng-Lin]的文章
百度学术
百度学术中相似的文章
[Yang, Peipei]的文章
[Huang, Kaizhu]的文章
[Liu, Cheng-Lin]的文章
必应学术
必应学术中相似的文章
[Yang, Peipei]的文章
[Huang, Kaizhu]的文章
[Liu, Cheng-Lin]的文章
相关权益政策
暂无数据
收藏/分享
文件名: chp%3A10.1007%2F978-3-642-33460-3_47.pdf
格式: Adobe PDF
此文件暂不支持浏览
所有评论 (0)
暂无评论
 

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。