Extensions of marginalized graph kernels

By Pierre Mahé, Nobuhisa Ueda, Tatsuya Akutsu, Jean-Luc Perret, Jean-Philippe Vert. 2004

224 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

How researchers see this paper

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

In Artificial Intelligence

In Computational Theory and Mathematics

Citations in the corpus

Text Categorization as a Graph Classification Problem (2015, background): “In parallel, the idea of marginalized kernels was extended to graphs by Kashima et al. (2003) and by Mahé et al. (2004).”
Graph Kernels (2008, background): “Much at the same time, the idea of marginalized kernels (Tsuda et al., 2002) was extended to graphs by Kashima et al. (2003, 2004), then further refined by Mahé et al. (2004).”
Graph kernels for chemical informatics (2005, background): “The corresponding results in Table 5 show, for example, that the MinMax kernel achieves a cross-validated performance accuracy of 91.5%, above all previously reported results [King et al., 1996, Mahé et al., 2004].”
Near-optimal Supervised Feature Selection among Frequent Subgraphs (2009, background): “Recent studies have proposed several effective graph classification methods, including substructure (or socalled fragment) based approaches [17, 8, 28], kernelbased approaches [16, 22, 12], and boosting methods [20].”
Fast subtree kernels on graphs (2009, background): “Mahé and Vert [9] define extensions of the classic subtree kernel that avoid tottering [8] and consider unbalanced subtrees.”
Ordered Decompositional DAG kernels enhancements (2015, background): “that compares random walks up to length p in two graphs (special case of [34] and [35] ) as representative for the family of kernels based on random walks, and the graphlet kernel [40].”
Generative Kernels for Tree-Structured Data (2018, methodology): “The marginalized kernel has been later extended to graphs [18], [36], [37], using a joint kernel that counts the number of matching pairs of random walks in two graphs, where the latent variable h is a sequence of graph vertices generated by a ﬁrst-order Markov random walk (as in HMM for sequences).”
Halting in Random Walk Kernels (2015, methodology): “5], a weakness of random walk kernels described more than a decade ago [11].”
Shortest-path kernels on graphs (2005, methodology): “One method to reach this is label enrichment as used in [16].”
Weisfeiler-Lehman Graph Kernels (2011, methodology): “Mahé et al. (2004) have proposed extensions of marginalized graph kernels (Kashima et al., 2003) for a chemoinformatics application: here the authors relabel vertices of graphs using the Morgan index (Morgan, 1965), which increases the specificity of labels by augmenting them with information on…”
The Graph Neural Network Model (2009, unknown): “For example, the diffusion kernel [33] is based on heat diffusion equation; the kernels proposed in [34] and [35] exploit the vectors produced by a graph random walker and those designed in [36]–[38] use a method of counting the number of common substructures of two trees.”
Marginalized Multi-Instance Kernels (2007, methodology): “This method has been successfully used to define marginalized kernels for strings, [Tsuda et al., 2002], trees and graphs [Mahé et al., 2004].”
Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity (2005, background): “Convolution graph kernels published in the literature (Gärtner et al., 2003; Kashima et al., 2003; Mahé et al., 2004) that can be applied to molecular graphs”