TY - JOUR
T1 - Online heterogeneous transfer by hedge ensemble of offline and online decisions
AU - Yan, Yuguang
AU - Wu, Qingyao
AU - Tan, Mingkui
AU - NG, Kwok Po
AU - Min, Huaqing
AU - Tsang, Ivor W.
N1 - Funding Information:
Manuscript received May 20, 2016; revised November 15, 2016, March 7, 2017, and August 31, 2017; accepted August 31, 2017. Date of publication October 10, 2017; date of current version June 21, 2018. This work was supported in part by the Guangzhou Key Laboratory of Robotics and Intelligent Software under Grant 15180007, in part by the National Natural Science Foundation of China under Grant 61502177 and Grant 61602185, in part by the Fundamental Research Funds for the Central Universities under Grant D2172500 and Grant D2172480, in part by the CCF-Tencent Open Research Fund, in part by the Guangdong Provincial Scientific and Technological Fund under Grant 2017B090901008 and Grant 2017A010101011, in part by HKRGC GRF under Grant 12302715, Grant 12306616, and Grant 12200317, in part by HKBU under Grant RC-ICRS/16-17/03, in part by the ARC Future Fellowship under Grant FT130100746, and in part by the ARC under Grant LP150100671. (Corresponding authors: Qingyao Wu; Mingkui Tan.) Y. Yan, Q. Wu, M. Tan, and H. Min are with the School of Software Engineering, South China University of Technology, Guangzhou 510006, China (e-mail: [email protected]; [email protected]; [email protected]; [email protected]).
PY - 2018/7
Y1 - 2018/7
N2 - In this paper, we study the online heterogeneous transfer (OHT) learning problem, where the target data of interest arrive in an online manner, while the source data and auxiliary co-occurrence data are from offline sources and can be easily annotated. OHT is very challenging, since the feature spaces of the source and target domains are different. To address this, we propose a novel technique called OHT by hedge ensemble by exploiting both offline knowledge and online knowledge of different domains. To this end, we build an offline decision function based on a heterogeneous similarity that is constructed using labeled source data and unlabeled auxiliary co-occurrence data. After that, an online decision function is learned from the target data. Last, we employ a hedge weighting strategy to combine the offline and online decision functions to exploit knowledge from the source and target domains of different feature spaces. We also provide a theoretical analysis regarding the mistake bounds of the proposed approach. Comprehensive experiments on three real-world data sets demonstrate the effectiveness of the proposed technique.
AB - In this paper, we study the online heterogeneous transfer (OHT) learning problem, where the target data of interest arrive in an online manner, while the source data and auxiliary co-occurrence data are from offline sources and can be easily annotated. OHT is very challenging, since the feature spaces of the source and target domains are different. To address this, we propose a novel technique called OHT by hedge ensemble by exploiting both offline knowledge and online knowledge of different domains. To this end, we build an offline decision function based on a heterogeneous similarity that is constructed using labeled source data and unlabeled auxiliary co-occurrence data. After that, an online decision function is learned from the target data. Last, we employ a hedge weighting strategy to combine the offline and online decision functions to exploit knowledge from the source and target domains of different feature spaces. We also provide a theoretical analysis regarding the mistake bounds of the proposed approach. Comprehensive experiments on three real-world data sets demonstrate the effectiveness of the proposed technique.
KW - Co-occurrence data
KW - hedge weighting
KW - heterogeneous transfer learning (HTL)
KW - online learning
UR - http://www.scopus.com/inward/record.url?scp=85031809891&partnerID=8YFLogxK
U2 - 10.1109/TNNLS.2017.2751102
DO - 10.1109/TNNLS.2017.2751102
M3 - Journal article
C2 - 29028211
AN - SCOPUS:85031809891
SN - 2162-237X
VL - 29
SP - 3252
EP - 3263
JO - IEEE Transactions on Neural Networks and Learning Systems
JF - IEEE Transactions on Neural Networks and Learning Systems
IS - 7
ER -