A1 Refereed original research article in a scientific journal
Limited memory bundle DC algorithm for sparse pairwise kernel learning
Authors: Karmitsa, Napsu; Joki, Kaisa; Airola, Antti; Pahikkala, Tapio
Publisher: Springer Science and Business Media LLC
Publication year: 2025
Journal: Journal of Global Optimization
Journal name in source: Journal of Global Optimization
Volume: 92
First page : 55
Last page: 85
ISSN: 0925-5001
eISSN: 1573-2916
DOI: https://doi.org/10.1007/s10898-025-01481-w
Web address : https://doi.org/10.1007/s10898-025-01481-w
Self-archived copy’s web address: https://research.utu.fi/converis/portal/detail/Publication/491873233
Pairwise learning is a specialized form of supervised learning that focuses on predicting outcomes for pairs of objects. In this paper, we formulate the pairwise learning problem as a difference of convex (DC) optimization problem using the Kronecker product kernel, ℓ1- and ℓ0-regularizations, and various, possibly nonsmooth, loss functions. Our aim is to develop an efficient learning algorithm, SparsePKL, that produces accurate predictions with the desired sparsity level. In addition, we propose a novel limited memory bundle DC algorithm (LMB-DCA) for large-scale nonsmooth DC optimization and apply it as an underlying solver in the SparsePKL. The performance of the SparsePKL-algorithm is studied in seven real-world drug-target interaction data and the results are compared with those of the state-of-art methods in pairwise learning.
Downloadable publication This is an electronic reprint of the original article. |  |
Funding information in the publication:
Open Access funding provided by University of Turku (including Turku University Central Hospital). Open Access funding provided by University of Turku (including Turku University Central Hospital. This work was financially supported by University of Turku and Research Council of Finland Grants #345804 and #345805.
