Issue No. 05 - May (2014 vol. 26)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TKDE.2013.111
Mingsheng Long , Sch. of Software, Tsinghua Univ., Beijing, China
Jianmin Wang , Sch. of Software, Tsinghua Univ., Beijing, China
Guiguang Ding , Sch. of Software, Tsinghua Univ., Beijing, China
Sinno Jialin Pan , Inst. of Infocomm Res., Singapore, Singapore
Philip S. Yu , Dept. of Comput. Sci., Univ. of Illinois at Chicago, Chicago, IL, USA
Domain transfer learning, which learns a target classifier using labeled data from a different distribution, has shown promising value in knowledge discovery yet still been a challenging problem. Most previous works designed adaptive classifiers by exploring two learning strategies independently: distribution adaptation and label propagation. In this paper, we propose a novel transfer learning framework, referred to as Adaptation Regularization based Transfer Learning (ARTL), to model them in a unified way based on the structural risk minimization principle and the regularization theory. Specifically, ARTL learns the adaptive classifier by simultaneously optimizing the structural risk functional, the joint distribution matching between domains, and the manifold consistency underlying marginal distribution. Based on the framework, we propose two novel methods using Regularized Least Squares (RLS) and Support Vector Machines (SVMs), respectively, and use the Representer theorem in reproducing kernel Hilbert space to derive corresponding solutions. Comprehensive experiments verify that ARTL can significantly outperform state-of-the-art learning methods on several public text and image datasets.
support vector machines, Hilbert spaces, learning (artificial intelligence), least squares approximations, minimisation, pattern classification
Mingsheng Long, Jianmin Wang, Guiguang Ding, S. J. Pan and P. S. Yu, "Adaptation Regularization: A General Framework for Transfer Learning," in IEEE Transactions on Knowledge & Data Engineering, vol. 26, no. 5, pp. 1076-1089, 2014.