This paper investigates a new machine learning framework of Online Transfer Learning (OTL), which aims to attack an online learning task on a target domain by transferring knowledge from some source domain. We do not assume data in the target domain follows the same distribution as that in the source domain, and the motivation of our work is to enhance a supervised online learning task on a target domain by exploiting the existing knowledge that had been learnt from training data in source domains. OTL is in general a challenging problem since data in both source and target domains not only can be different in their class distributions, but also can be different in their feature representations. As a first attempt to this new research, we investigate two different settings of OTL: (i) OTL on homogeneous domains of common feature space, and (ii) OTL across heterogeneous domains of different feature spaces. For each setting, we propose OTL algorithms to solve two tasks: classification and regression, and show the theoretical bounds of the proposed algorithms. In addition, we also apply the OTL technique to solve the concept-drifting data stream learning problem, a real-life challenge in data mining and machine learning. Finally, we conduct extensive empirical studies on a comprehensive testbed, in which encouraging results validate the efficacy of our techniques.
Transfer learning, Online learning, Knowledge transfer
Databases and Information Systems
Data Management and Analytics
ZHAO, Peilin; HOI, Steven C. H.; WANG, Jialei; and LI, Bin.
Online Transfer Learning. (2014). Artificial Intelligence. 216, 76-102. Research Collection School Of Information Systems.
Available at: http://ink.library.smu.edu.sg/sis_research/2269
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.