Home > Research > Publications & Outputs > A Generalized Weisfeiler-Lehman Graph Kernel

Electronic data

  • 2101.08104v1

    Submitted manuscript, 426 KB, PDF document

Links

Keywords

View graph of relations

A Generalized Weisfeiler-Lehman Graph Kernel

Research output: Working paperPreprint

Published

Standard

A Generalized Weisfeiler-Lehman Graph Kernel. / Schulz, Till Hendrik; Horváth, Tamás; Welke, Pascal et al.
Arxiv, 2021.

Research output: Working paperPreprint

Harvard

APA

Vancouver

Schulz TH, Horváth T, Welke P, Wrobel S. A Generalized Weisfeiler-Lehman Graph Kernel. Arxiv. 2021 Jan 20.

Author

Schulz, Till Hendrik ; Horváth, Tamás ; Welke, Pascal et al. / A Generalized Weisfeiler-Lehman Graph Kernel. Arxiv, 2021.

Bibtex

@techreport{25089fdc33cd48949d2969f14f6ce59f,
title = "A Generalized Weisfeiler-Lehman Graph Kernel",
abstract = " The Weisfeiler-Lehman graph kernels are among the most prevalent graph kernels due to their remarkable time complexity and predictive performance. Their key concept is based on an implicit comparison of neighborhood representing trees with respect to equality (i.e., isomorphism). This binary valued comparison is, however, arguably too rigid for defining suitable similarity measures over graphs. To overcome this limitation, we propose a generalization of Weisfeiler-Lehman graph kernels which takes into account the similarity between trees rather than equality. We achieve this using a specifically fitted variation of the well-known tree edit distance which can efficiently be calculated. We empirically show that our approach significantly outperforms state-of-the-art methods in terms of predictive performance on datasets containing structurally more complex graphs beyond the typically considered molecular graphs. ",
keywords = "cs.LG",
author = "Schulz, {Till Hendrik} and Tam{\'a}s Horv{\'a}th and Pascal Welke and Stefan Wrobel",
note = "n/a",
year = "2021",
month = jan,
day = "20",
language = "English",
publisher = "Arxiv",
type = "WorkingPaper",
institution = "Arxiv",

}

RIS

TY - UNPB

T1 - A Generalized Weisfeiler-Lehman Graph Kernel

AU - Schulz, Till Hendrik

AU - Horváth, Tamás

AU - Welke, Pascal

AU - Wrobel, Stefan

N1 - n/a

PY - 2021/1/20

Y1 - 2021/1/20

N2 - The Weisfeiler-Lehman graph kernels are among the most prevalent graph kernels due to their remarkable time complexity and predictive performance. Their key concept is based on an implicit comparison of neighborhood representing trees with respect to equality (i.e., isomorphism). This binary valued comparison is, however, arguably too rigid for defining suitable similarity measures over graphs. To overcome this limitation, we propose a generalization of Weisfeiler-Lehman graph kernels which takes into account the similarity between trees rather than equality. We achieve this using a specifically fitted variation of the well-known tree edit distance which can efficiently be calculated. We empirically show that our approach significantly outperforms state-of-the-art methods in terms of predictive performance on datasets containing structurally more complex graphs beyond the typically considered molecular graphs.

AB - The Weisfeiler-Lehman graph kernels are among the most prevalent graph kernels due to their remarkable time complexity and predictive performance. Their key concept is based on an implicit comparison of neighborhood representing trees with respect to equality (i.e., isomorphism). This binary valued comparison is, however, arguably too rigid for defining suitable similarity measures over graphs. To overcome this limitation, we propose a generalization of Weisfeiler-Lehman graph kernels which takes into account the similarity between trees rather than equality. We achieve this using a specifically fitted variation of the well-known tree edit distance which can efficiently be calculated. We empirically show that our approach significantly outperforms state-of-the-art methods in terms of predictive performance on datasets containing structurally more complex graphs beyond the typically considered molecular graphs.

KW - cs.LG

M3 - Preprint

BT - A Generalized Weisfeiler-Lehman Graph Kernel

PB - Arxiv

ER -