arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1608.04581cs.LGstat.ML
收藏

基于公共空间映射和加权域匹配的迁移学习新方法

A novel transfer learning method based on common space mapping and weighted domain matching

King Abdullah University of Science and Technology（阿卜杜拉国王科技大学）；Vanderbilt University（范德堡大学）；Suning Commerce R&D Center USA, Inc（苏宁美国研发中心）；Harbin University of Science and Technology（哈尔滨理工大学）；New York University Abu Dhabi（纽约大学阿布扎比分校）；Michigan State University（密歇根州立大学）

Ru-Ze Liang, Wei Xie, Weizhi Li, Hongqi Wang, Jim Jing-Yan Wang, Lisa Taylor

备注：
arXiv admin note: text overlap with arXiv:1605.06673
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

摘要：

针对领域迁移学习问题，提出了一种新的学习框架。我们将两个域的数据映射到一个公共空间，并在这个公共空间中学习分类器。然后，我们适应共同的分类器，这两个领域分别增加了两个自适应功能。在公共空间中，对目标域数据点进行加权，并根据分布匹配到目标域。源域数据点和目标域分类响应的加权项也由局部重构系数正则化。新的迁移学习框架在一些基准跨域数据集上进行了评估，它优于现有的最先进的迁移学习方法。

英文摘要：

In this paper, we propose a novel learning framework for the problem of domain transfer learning. We map the data of two domains to one single common space, and learn a classifier in this common space. Then we adapt the common classifier to the two domains by adding two adaptive functions to it respectively. In the common space, the target domain data points are weighted and matched to the target domain in term of distributions. The weighting terms of source domain data points and the target domain classification responses are also regularized by the local reconstruction coefficients. The novel transfer learning framework is evaluated over some benchmark cross-domain data sets, and it outperforms the existing state-of-the-art transfer learning methods.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑