Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Online Ensemble Multi-kernel Learning Adaptive to Non-stationary and Adversarial Environments
(Submitted on 28 Dec 2017 (this version), latest version 28 Dec 2018 (v3))
Abstract: Kernel-based methods exhibit well-documented performance in various nonlinear learning tasks. Most of them rely on a preselected kernel, whose prudent choice presumes task-specific prior information. To cope with this limitation, multi-kernel learning has gained popularity thanks to its flexibility in choosing kernels from a prescribed kernel dictionary. Leveraging the random feature approximation and its recent orthogonality-promoting variant, the present contribution develops an online multi-kernel learning scheme to infer the intended nonlinear function `on the fly.' To further boost performance in non-stationary environments, an adaptive multi-kernel learning scheme (termed AdaRaker) is developed with affordable computation and memory complexity. Performance is analyzed in terms of both static and dynamic regret. To our best knowledge, AdaRaker is the first algorithm that can optimally track nonlinear functions in non-stationary settings with theoretical guarantees. Numerical tests on real datasets are carried out to showcase the effectiveness of the proposed algorithms.
Submission history
From: Yanning Shen [view email][v1] Thu, 28 Dec 2017 18:42:38 GMT (1028kb,D)
[v2] Wed, 17 Jan 2018 05:11:34 GMT (3030kb,D)
[v3] Fri, 28 Dec 2018 06:33:22 GMT (1304kb,D)
Link back to: arXiv, form interface, contact.