Adaptive learning in a world of projections

Sergios Theodoridis, Konstantinos Slavakis, Isao Yamada

    Research output: Contribution to journalArticlepeer-review

    162 Scopus citations

    Abstract

    This article presents a general tool for convexly constrained parameter/function estimation both for classification and regression tasks, in a time-adaptive setting and in (infinite dimensional) Reproducing Kernel Hilbert Spaces (RKHS). The mathematical framework is that of the set theoretic estimation formulation and the classical projections onto convex sets (POCS) theory. However, in contrast to the classical POCS methodology, which assumes a finite number of convex sets, our method builds upon our recent extension of the theory, which considers an infinite number of convex sets. Such a context is necessary to cope with the adaptive setting rationale, where data arrive sequentially.

    Original languageEnglish (US)
    Article number5670637
    Pages (from-to)97-123
    Number of pages27
    JournalIEEE Signal Processing Magazine
    Volume28
    Issue number1
    DOIs
    StatePublished - Jan 2011

    Fingerprint Dive into the research topics of 'Adaptive learning in a world of projections'. Together they form a unique fingerprint.

    Cite this