Stream-Based Active Learning with Multiple Kernels

Abstract

Online multiple kernel learning (OMKL) has provided an attractive performance in nonlinear function learning tasks. Leveraging a random feature (RF) approximation, the major drawback of OMKL, known as the curse of dimensionality, has been recently alleviated. These advantages enable RF-based OMKL to be considered in practice. In this paper we introduce a new research problem, named stream-based active multiple kernel learning (AMKL), where a learner is allowed to label some selected data from an oracle according to a selection criterion. This is necessary in many real-world applications since acquiring a true label is costly or time-consuming. We theoretically prove that the proposed AMKL achieves an optimal sublinear regret \mathcal{O}(\sqrt{T}) as in OMKL with little labeled data, implying that the proposed selection criterion indeed avoids unnecessary label-requests.