Large Scale Multiple Kernel Learning

By Sören Sonnenburg, Gunnar Rätsch, Christin Schäfer, Bernhard Schölkopf. 2006

1,427 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

Bayesian Learning of Kernel Embeddings (2016, methodology): “…is an effective and widely used method for kernel selection, and the myriad papers on multiple kernel learning (e.g. Bach et al. (2004); Sonnenburg et al. (2006); G¨onen and Alpaydın (2011)) assume that some loss function is available and thus focus on effective ways of learning…”
A Survey on Multi-view Learning (2013, methodology): “This novel multiple kernel learning framework is called simple MKL, which has been shown to be more efficient than the SILP problem.”
Optimization with Sparsity-Inducing Penalties (2011, methodology): “This is also the case for algorithms designed for classical multiple kernel learning when the regularizer is the squared norm [111, 129, 132]; these methods are therefore not exactly comparable to the MKL algorithms presented in this monograph which apply to objective regularized by the unsquared no”
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).”
Factorized Orthogonal Latent Spaces (2010, background): “…in real scenarios, information in the views is typically neither fully independent nor fully correlated, and thus multi-kernel (Bach et al., 2004; Sonnenburg et al., 2006) and shared latent space (Kuss and Grae-pel, 2003; Ek et al., 2007; Navaratnam et al., 2007; Si-gal et al., 2009) techniques…”
Factorized Latent Spaces with Structured Sparsity (2010, background): “Multiple kernel learning [2, 24] methods have proven successful under the assumption that the views are independent.”
Legal, regulatory, and ethical frameworks for development of standards in artificial intelligence (AI) and autonomous robotic surgery (2019, background): “30 However, in pathology, we are often confronted with bias in data sets, and spurious patterns in data derived from affluent/racially/gendered, non ‐ diverse samples.”
Online Learning: A Comprehensive Survey (2018, methodology): “Specifically, the OMKL in (Jin et al., 2010; Hoi et al., 2013) learns a kernel-based prediction function by selecting a subset of predefined kernel functions in an online learning fashion, which is in general more challenging than typical online learning because both the kernel classifiers and the s”
Semi-Supervised Learning via Generalized Maximum Entropy (2010, unknown): “U = ‖ Ep[ψ]p− ψ̃ Ep[φ]p− 0  ‖H where H = (H1,H2) with kernel K = K1 +K2 (Bach et al., 2004; Sonnenburg et al., 2006).”
Kernel Methods in Computer Vision (2009, methodology): “The resulting optimization problem is jointly-convex in the vectors vj and in the kernel coefficients βj [Sonnenburg et al., 2006].”
Supervised classification and mathematical optimization (2013, methodology): “For instance, for K taken as the set of convex combinations of m given kernels, [10] gives an adaptation of SMO, while in [217] the problem is formulated as a semi-infinite linear program and solved via column generation.”
Concurrent Action Detection with Structural Prediction (2013, methodology): “We visualizes the learned weights of human body joints (the normalized weights of multiple kernels [15]) in Figure 6.”
Predicting abnormal returns from news using text classification (2008, methodology): “vely optimizes over pairs of variables is used to solve problem (4). Other approaches for solving larger scale problems are written as a wrapper around an SVM computation. For example, an approach in [5] was to solve the semi-inﬁnite linear program (SILP) formulation: maximize λ subject to P i di = ”
Learning to predict where humans look (2009, methodology): “We used models with linear kernels because we found from experimentation that they performed as well as models with radial basis function kernels and models found with multiple kernel learning [18] for our speciﬁc task.”
Localized multiple kernel learning (2008, methodology): “In recent studies (Lanckriet et al., 2004a; Sonnenburg et al., 2006), it is reported that using multiple di.erent kernels instead of a single kernel improves the classi.”
Multiple Operator-valued Kernel Learning (2012, background): “In order to overcome the need for choosing a kernel before the learning process, several works have tried to address the problem of learning the scalar-valued kernel jointly with the decision function [18, 29].”
A glass-box interactive machine learning approach for solving NP-hard problems with the human-in-the-loop (2017, background): “Consequently, automatic machine learning (aML) works well when having large amounts of training data (Sonnenburg et al., 2006), consequently “big data”, which is often considered as burden - is here not only beneﬁcial but necessary.”