Efficient and Accurate Lp-Norm Multiple Kernel Learning

By Marius Kloft, Ulf Brefeld, Sören Sonnenburg, Pavel Laskov, K. Müller, Alexander Zien. 2009

260 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep multilayer multiple kernel learning (2015, methodology): “During the last decade, MKL has been actively investigated, in which lots of algorithms and techniques were proposed in order to improve the efficiency of the MKL method [8, 13, 19, 22, 24, 27, 28, 30].”
Simple and Efficient Multiple Kernel Learning by Group Lasso (2010, methodology): “The advantage of the alternating scheme is that SVM solvers can be very efficient due to recent advances in large scale optimization (Bottou & Lin, 2007).”
Kernel-Based Weighted Multi-view Clustering (2012, methodology): “A weighted combination of the kernels that resembles the p-norm regularizer [22] and reflects the views’ relevance to the clustering task is automatically learned using closed form updates.”
$\ell_{p}-\ell_{q}$ Penalty for Sparse Linear and Sparse Multiple Kernel Multitask Learning (2011, methodology): “However, according to very recent works on MKL [ 25 ], [46], using such a wrapper approach (which first solves an SVM then update the weights d) is still competitive compared to other algorithms.”
Compositional Models for Video Event Detection: A Multiple Kernel Learning Latent Variable Approach (2013, methodology): “The kernel weights can be updated using a Newton descent step or the cutting plane approach [5].”
Multi-Task Multiple Kernel Relationship Learning (2016, methodology): “This can be thought of as a simple extension of lp-MKL to multi-task setting [11].”
Data Fusion by Matrix Factorization (2013, methodology): “Extension from single to multiple data sources is achieved by additive combination of kernel matrices, given by V 1⁄4 Pri1⁄41 uiKi j 8i : ui 0; Pri1⁄41 udi 1⁄4 1;Ki 0g; where ui are weights of the kernel matrices, d is a parameter determining the norm of constraint posed on coefficients (for L2; Lp-”
Multiclass transfer learning from unconstrained priors (2011, background): “Recently, it has been extended to lp norm regularization in [16, 21] for tuning the level of sparsity with the additional parameter p.”
Efficient Rule Ensemble Learning using Hierarchical Kernels (2011, methodology): “, ρ ∈ (1, 4/3)) (Szafranski et al., 2008; Kloft et al., 2009).”
Inferring latent task structure for Multitask Learning by Multiple Kernel Learning (2010, methodology): “For an in depth discussion of this, please consider [19].”
Multiple Kernel Learning from Noisy Labels by Stochastic Programming (2012, methodology): “, 2010) optimize the same objective as in (2), (ii) some algorithms (Cortes et al., 2009; Kloft et al., 2009) focus on different regularizations (e.”
Insights from Classifying Visual Concepts with Multiple Kernel Learning (2011, methodology): “While prior work on MKL imposes a1-norm constraint on the mixing coefficients to enforce sparse solutions lying on a standard simplex [14, 15, 32, 33], we employ a generalizedlp-norm constraint‖β‖p ≤ 1 for p ≥ 1 as used in [20, 21].”
A Convex Feature Learning Formulation for Latent Task Structure Discovery (2012, unknown): “In the special case p̄ = q̄, this problem is same as the `q̂-MKL formulation (Kloft et al., 2009) with q̂ = q̄q̄−1 and with base kernels as k̂jw = (λw(γ)) 1 q̄ kjw ∀w ∈ V and ∀j = 1, . . . , n.”