OTL: A Framework of Online Transfer Learning
Peilin Zhao, Steven C. H. Hoi · 2010
In this paper, we investigate a new machine learning framework called Online Transfer Learning (OTL) that aims to transfer knowl-edge from some source domain to an online learning task on a target domain. We do not assume the target data follows the same class or generative distribution as the source data, and our key motivation is to improve a su-pervised online learning task in a target do-main by exploiting the knowledge that had been learned from large amount of training data in source domains. OTL is in general challenging since data in both domains not only can be different in their class distribu-tions but can be also different in their fea-ture representations. As a first attempt to this problem, we propose techniques to ad-dress two kinds of OTL tasks: one is to per-form OTL in a homogeneous domain, and the other is to perform OTL across heteroge-neous domains. We show the mistake bounds of the proposed OTL algorithms, and empir-ically examine their performance on several challenging OTL tasks. Encouraging results validate the efficacy of our techniques. 1.