###
Journal of Software:2015.26(5):977-1000

稀疏标签传播:一种鲁棒的领域适应学习方法
陶剑文,Fu-LaiCHUNG,王士同,姚奇富
(浙江大学 宁波理工学院 信息科学与工程学院, 浙江 宁波 315100;香港理工大学 电子计算学系, 香港;香港理工大学 电子计算学系, 香港;江南大学 数字媒体学院, 江苏 无锡 214122;浙江工商职业技术学院 电子与信息工程学院, 浙江 宁波 315012)
Sparse Label Propagation: A Robust Domain Adaptation Learning Method
TAO Jian-Wen,Fu-Lai CHUNG,WANG Shi-Tong,YAO Qi-Fu
(School of Information Science and Engineering, Ningbo Institute of Technology, Zhejiang University, Ningbo 315100, China;Department of Computing, Hong Kong Polytechnic University, Hong Kong, China;Department of Computing, Hong Kong Polytechnic University, Hong Kong, China;School of Digital Media, Jiangnan University, Wuxi 214122, China;School of Information Engineering, Zhejiang Business Technology Institute, Ningbo 315012, China)
Abstract
Chart / table
Reference
Similar Articles
Article :Browse 2913   Download 3137
Received:January 21, 2013    Revised:January 10, 2014
> 中文摘要: 稀疏表示因其所具有的鲁棒性,在模式分类领域逐渐得到关注.研究了一种基于稀疏保留模型的新颖领域适应学习方法,并提出一种鲁棒的稀疏标签传播领域适应学习(sparse label propagation domain adaptation learning,简称SLPDAL)算法.SLPDAL通过将目标领域数据进行稀疏重构,以实现源领域数据标签向目标领域平滑传播.具体来讲,SLPDAL算法分为3步:首先,基于领域间数据分布均值差最小化准则寻求一个优化的核空间,并将领域数据嵌入到该核空间;然后,在该嵌入核空间,基于l1-范最小化准则计算各领域数据的核稀疏重构系数;最后,通过保留领域数据间核稀疏重构系数约束,实现源领域数据标签向目标领域的传播.最后,将SLPDAL算法推广到多核学习框架,提出一个SLPDAL多核学习模型.在鲁棒人脸识别、视频概念检测和文本分类等领域适应学习任务上进行比较实验,所提出的方法取得了优于或可比较的学习性能.
Abstract:Sparse representation has received an increasing amount of interest in pattern classification due to its robustness. In this paper, a domain adaptation learning (DAL) approach is explored based on a sparsity preserving model, which assumes that each data point can be sparsely reconstructed. The proposed robust DAL algorithm, called sparse label propagation domain adaptation learning (SLPDAL), propagates the labels from labeled points in the source domain to the unlabeled dataset in the target domain using those sparsely reconstructed objects with sufficient smoothness. SLPDAL consists of three steps. First, it finds an optimal kernel space in which all samples from both source and target domains can be embedded by minimizing the mean discrepancy between these two domains. Then, it computes the best kernel sparse reconstructed coefficients for each data point in the kernel space by using l1-norm minimization. Finally, it propagates the labels of source domain to the target domain by preserving the kernel sparse reconstructed coefficients. The paper also derives an easy way to extend SLPDAL to out-of-sample data and multiple kernel learning respectively. Promising experimental results have been obtained for several DAL problems such as face recognition, visual video detection and text classification tasks.
文章编号:     中图分类号:    文献标志码:
基金项目:教育部人文社会科学研究规划基金(13YJAZH084); 浙江省自然科学基金(LY14F020009); 宁波市自然科学基金(2013A610065, 2013A610072); 香港理工大学基金(G-UA68) 教育部人文社会科学研究规划基金(13YJAZH084); 浙江省自然科学基金(LY14F020009); 宁波市自然科学基金(2013A610065, 2013A610072); 香港理工大学基金(G-UA68)
Foundation items:
Reference text:

陶剑文,Fu-Lai CHUNG,王士同,姚奇富.稀疏标签传播:一种鲁棒的领域适应学习方法.软件学报,2015,26(5):977-1000

TAO Jian-Wen,Fu-Lai CHUNG,WANG Shi-Tong,YAO Qi-Fu.Sparse Label Propagation: A Robust Domain Adaptation Learning Method.Journal of Software,2015,26(5):977-1000