lp-Norm Multiple Kernel Learning

Research output: Journal contributionsJournal articlesResearchpeer-review

Standard

lp-Norm Multiple Kernel Learning. / Kloft, Marius; Brefeld, Ulf; Sonnenburg, Sören et al.
In: Journal of Machine Learning Research, Vol. 2011, No. 12, 2011, p. 953-997.

Research output: Journal contributionsJournal articlesResearchpeer-review

Harvard

APA

Vancouver

Kloft M, Brefeld U, Sonnenburg S, Zien A. lp-Norm Multiple Kernel Learning. Journal of Machine Learning Research. 2011;2011(12):953-997.

Bibtex

@article{1c2221df8dec4ba085cf168bdcc8fe59,
title = "lp-Norm Multiple Kernel Learning",
abstract = "Learning linear combinations of multiple kernels is an appealing strategy when the right choice of features is unknown. Previous approaches to multiple kernel learning (MKL) promote sparse kernel combinations to support interpretability and scalability. Unfortunately, this `1-norm MKL is rarely observed to outperform trivial baselines in practical applications. To allow for robust kernelmixtures that generalize well, we extend MKL to arbitrary norms. We devise new insights on the connection between several existing MKL formulations and develop two efficient interleaved optimization strategies for arbitrary norms, that is `p-norms with p1. This interleaved optimization is much faster than the commonly used wrapper approaches, as demonstrated on several data sets. A theoretical analysis and an experiment on controlled artificial data shed light on the appropriateness of sparse, non-sparse and `¥-norm MKL in various scenarios. Importantly, empirical applications of `p-norm MKL to three real-world problems from computational biology show that non-sparse MKL achieves accuracies that surpass the state-of-the-art.",
keywords = "Informatik, Wirtschaftsinformatik",
author = "Marius Kloft and Ulf Brefeld and S{\"o}ren Sonnenburg and Alexander Zien",
year = "2011",
language = "Deutsch",
volume = "2011",
pages = "953--997",
journal = "Journal of Machine Learning Research",
issn = "1532-4435",
publisher = "Microtome Publishing",
number = "12",

}

RIS

TY - JOUR

T1 - lp-Norm Multiple Kernel Learning

AU - Kloft, Marius

AU - Brefeld, Ulf

AU - Sonnenburg, Sören

AU - Zien, Alexander

PY - 2011

Y1 - 2011

N2 - Learning linear combinations of multiple kernels is an appealing strategy when the right choice of features is unknown. Previous approaches to multiple kernel learning (MKL) promote sparse kernel combinations to support interpretability and scalability. Unfortunately, this `1-norm MKL is rarely observed to outperform trivial baselines in practical applications. To allow for robust kernelmixtures that generalize well, we extend MKL to arbitrary norms. We devise new insights on the connection between several existing MKL formulations and develop two efficient interleaved optimization strategies for arbitrary norms, that is `p-norms with p1. This interleaved optimization is much faster than the commonly used wrapper approaches, as demonstrated on several data sets. A theoretical analysis and an experiment on controlled artificial data shed light on the appropriateness of sparse, non-sparse and `¥-norm MKL in various scenarios. Importantly, empirical applications of `p-norm MKL to three real-world problems from computational biology show that non-sparse MKL achieves accuracies that surpass the state-of-the-art.

AB - Learning linear combinations of multiple kernels is an appealing strategy when the right choice of features is unknown. Previous approaches to multiple kernel learning (MKL) promote sparse kernel combinations to support interpretability and scalability. Unfortunately, this `1-norm MKL is rarely observed to outperform trivial baselines in practical applications. To allow for robust kernelmixtures that generalize well, we extend MKL to arbitrary norms. We devise new insights on the connection between several existing MKL formulations and develop two efficient interleaved optimization strategies for arbitrary norms, that is `p-norms with p1. This interleaved optimization is much faster than the commonly used wrapper approaches, as demonstrated on several data sets. A theoretical analysis and an experiment on controlled artificial data shed light on the appropriateness of sparse, non-sparse and `¥-norm MKL in various scenarios. Importantly, empirical applications of `p-norm MKL to three real-world problems from computational biology show that non-sparse MKL achieves accuracies that surpass the state-of-the-art.

KW - Informatik

KW - Wirtschaftsinformatik

M3 - Zeitschriftenaufsätze

VL - 2011

SP - 953

EP - 997

JO - Journal of Machine Learning Research

JF - Journal of Machine Learning Research

SN - 1532-4435

IS - 12

ER -

Recently viewed

Publications

  1. Comparison of Bio-Inspired Algorithms in a Case Study for Optimizing Capacitor Bank Allocation in Electrical Power Distribution
  2. Tuning kalman filter in linear systems
  3. How mobile app design impacts user responses to mixed self-tracking outcomes
  4. Implementation of formative assessment
  5. A Comparative Study for Fisheye Image Classification
  6. The role of task meaning on output in groups
  7. Optimal scheduling of AGVs in a reentrant blocking job-shop
  8. General Patterns and Conclusions
  9. Model-based nonlinear filter design for tower load reduction of wind power plants with active power control capability
  10. Self-perceived quality of life predicts mortality risk better than a multi-biomarker panel, but the combination of both does best
  11. Comparison of Supervised versus Self-Administered Stretching on Bench Press Maximal Strength and Force Development
  12. Putting adaptive planning into practice: A meta-analysis of current applications
  13. Almost-invariant and finite-time coherent sets
  14. A common European asylum system? How variation in Member States’ administrative capacity undermines EU asylum harmonisation
  15. Studying properties of water data using manifold-aware anomaly detectors
  16. Algorithmisches Management
  17. Quantification of amino acids in fermentation media by isocratic HPLC analysis of their
  18. A Stacked Planar Sensor Concept for Minimally Invasive Plasma Monitoring
  19. How to move the transition to sustainable food consumption towards a societal tipping point
  20. A review of mobile language learning applications
  21. Microsimulation - A survey of principles, developments and applications