Marginalized kernels for biological sequences

By Koji Tsuda, Taishin Kin, Kiyoshi Asai. 2002

208 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Molecular Biology

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Molecular Biology

Citations in the corpus

Fastfood: Approximate Kernel Expansions in Loglinear Time (2014, methodology): “A related strategy can be found in the so-called ‘empirical’ kernel map (Tsuda et al., 2002; Sch¨olkopf and Smola, 2002) where kernels are computed via for x i often drawn from the same distribution as the training data.”
Generative and Latent Mean Map Kernels (2010, background): “The expectation of a feature map is similar to the marginalized kernel [26], although that work does not discuss RKHSs and focuses on count kernels rather than higher order (Taylor type) kernels such as the Gaussian RBF kernel.”
Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference (2015, background): “It is straightforward to show that K ( x , y ) is a valid PSD covariance function – it is an example of a marginalised covariance function [12].”
Graph Kernels (2008, neutral): “Much at the same time, the idea of marginalized kernels (Tsuda et al., 2002) was extended to graphs by Kashima et al. (2003, 2004), and further refined by Mahé et al. (2004).”
Protein-ligand interaction prediction: an improved chemogenomics approach (2008, background): “Depending on the target families, various categories of compounds are defined to indicate the type of interaction between each target and each compound.”
Extensions of marginalized graph kernels (2004, background): “Equation (1) can be seen as a marginalized kernel (Tsuda et al., 2002) and is therefore a valid kernel on the set of graphs.”
Dropout as a Bayesian Approximation : Insights and Applications (2015, background): “It is trivial to show that this defines a valid covariance function following (Tsuda et al., 2002).”
Generative Kernels for Tree-Structured Data (2018, background): “The marginalized kernel for sequences [35], in particular, exploits HMMs to define a joint kernel that counts the co-occurrences of hidden states and observed labels in the sequences and weights them by the posterior probabili-”
Cluster Kernels for Semi-Supervised Learning (2002, background): “Independently, another extension of the Fisher kernel has been proposed in [12] which leads, in the case of a mixture of Gaussians (J.”
Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning (2015, methodology): “ some element-wise non-linearity ˙() and distributions p(w);p(b). It is straightforward to show that K(x;y) is a valid PSD covariance function – it is an example of a marginalised covariance function [28]. In section 3 in the appendix we show that a deep Gaussian process with Llayers and covariance ”
Marginalized Multi-Instance Kernels (2007, methodology): “As mentioned in [Tsuda et al., 2002], an advantage of the marginalized kernel approach is that the definitions of the joint kernel and probabilistic model are completely separated.”
Self-taught learning: transfer learning from unlabeled data (2007, methodology): “In contrast, self-taught learning 13 In our experiments, the marginalized kernel (Tsuda et al., 2002), that takes an expectation over a (computed by MCMC sampling) rather than the MAP approximation, did not perform better.”
Genomic Similarity and Kernel Methods II: Methods for Genomic Information (2010, unknown): “An alternative probabilistic approach, called marginalized kernels [11] , uses the expectation of a kernel, with the expectation taken over unobserved latent variables.”
Support Vector Machines and Kernels for Computational Biology (2008, methodology): “Other kernels based on probabilistic models include the Covariance kernel [40] and Marginalized kernels [41].”