Deep Graph Kernels

By Pinar Yanardag, Pinar Yanardag, S. V. N. Vishwanathan. 2015

1,512 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Graph Learning

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 Graph Learning

Citations in the corpus

Diffusion-Convolutional Neural Networks (2015, background): “Deep graph kernels decompose a graph into substructures, treat those substructures as words in a sentence, and fit a word-embedding model to obtain a vectorization [6].”
Auto-Keras: An Efficient Neural Architecture Search System (2018, methodology): “Inspired by Deep Graph Kernels [38], we propose an edit-distance kernel for neural networks.”
The More You Know: Using Knowledge Graphs for Image Classification (2016, background): “is also a substantial amount of work on various types of kernels deﬁned for graphs [36] such as diffusion kernels [14], graphlet kernels [33], Weisfeiler-Lehman graph kernels [32], deep graph kernels [27], graph invariant kernels [26] and shortest-path kernels [1]. The methods have various ways of e”
Position-aware Graph Neural Networks (2019, background): “Weisfeiler-Lehman and Subgraph kernels have been incorporated into deep graph kernels (Yanardag & Vishwanathan, 2015) to capture structural properties of neighborhoods.”
Unsupervised Inductive Whole-Graph Embedding by Preserving Graph Proximity (2019, methodology): “The three deep graph kernels, i.e. DGK, DSP, and WDL, generate the same embeddings as the non-deep versions, but use additional techniques Yanardag & Vishwanathan (2015) to modify the similarity kernel matrices, resulting in different classification and ranking performance.”
Unsupervised Inductive Graph-Level Representation Learning via Graph-Graph Proximity (2019, methodology): “However, most existing methods for graph-level embeddings assume a supervised model [Ying et al., 2018; Zhang and Chen, 2019], with only a few exceptions, such as GRAPH KERNELS [Yanardag and Vishwanathan, 2015] and GRAPH2VEC [Narayanan et al., 2017].”
How Powerful are Graph Neural Networks? (2018, background): “Capturing structural similarity of the node labels is shown to be helpful for generalization particularly when the co-occurrence of subtrees is sparse across different graphs or there are noisy edges and node features (Yanardag & Vishwanathan, 2015).”
An End-to-End Deep Learning Architecture for Graph Classification (2018, methodology): “Datasets We compare DGCNN with other deep learning approaches for graph classification on six datasets, including three benchmark bioinformatics datasets: NCI1, PROTEINS, and D&D, as well as three social network datasets: COLLAB, IMDB-B, IMDB-M (Yanardag and Vishwanathan 2015).”
subgraph2vec: Learning Distributed Representations of Rooted Sub-graphs from Large Graphs (2016, methodology): “This will produce a vocabulary of subgraphs, SGvocab = {sg1, sg2, ...} from all the graphs in G δ: number of dimensions (embedding size) e: number of epochs output: Matrix of vector representations of subgraphs Φ ∈ R|SGvocab|×δ 1 begin 2 SGvocab = BuildSubgraphVocab(G) //use Algorithm 2 3…”
Graphite: Iterative Generative Modeling of Graphs (2018, methodology): “ed based on graph properties such as subgraph patterns (Shervashidze et al., 2009; Shervashidze and Borgwardt, 2009) and shortest paths (Borgwardt and Kriegel, 2005), or explicitly learned from data (Yanardag and Vishwanathan, 2015). See Vishwanathan et al. (2010) for a detailed review. In fact, the”
Provably Powerful Graph Networks (2019, methodology): “For classiﬁcation, we tested our method on eight real-world graph datasets from (Yanardag and Vishwanathan, 2015): three datasets consist of social network graphs, and the other ﬁve datasets come from bioinformatics and represent chemical compounds or protein structures.”
Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks (2018, methodology): “Graphlet and Weisfeiler-Leman kernels have been successfully employed within frameworks for smoothed and deep graph kernels (Yanardag and Vishwanathan 2015a; Yanardag and Vishwanathan 2015b).”
A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications (2017, methodology): “Yanardag and Vishwanathan [93] learn the embedding for subgraphs (e.g., graphlets) so as to define the graph kernels for graph classification.”
Simplifying Graph Convolutional Networks (2019, methodology): “We replace the GCN in DCGCN (Zhang et al., 2018b) with an SGC and get 71.0% and 76.2% on NCI1 and COLLAB datasets (Yanardag & Vishwanathan, 2015) respectively, which is on par with an GCN counterpart, but far behind GIN.”
A simple yet effective baseline for non-attribute graph classification (2018, background): “…our baseline with 6 state-of-the-art graph kernels: Weisfeiler-Lehman Kernel (WL)[SSL + 11], Graphlet kernel (GK)[SVP + 09], Deep Graph Kernel (DGK)[YV15a], RetGK[ZWX + 18], FGSD[VZ17], Weisfeiler-Lehman optimal assignment kernel (WL-OA)[KGW16], and 5 graph neural networks: PATCHYSAN (PSCN)…”
Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs (2017, neutral): “[34] and to four approaches using deep learning as at least one of their components [2, 27, 40, 8].”
On Valid Optimal Assignment Kernels and Applications to Graph Classification (2016, background): “[4, 23, 19, 24]: MUTAG, PTC-MR, NCI1 and NCI109 are graphs derived from small molecules, PROTEINS, D&D and ENZYMES represent macromolecules, and COLLAB and REDDIT are derived from social networks.”
graph2vec: Learning Distributed Representations of Graphs (2017, methodology): “DOI: 10.1145/1235 malware [6] and those of chemical compounds could be used to predict their properties such as solubility and anti-cancer activity [7].”
Walklets: Multiscale Graph Embeddings for Interpretable Network Classification (2016, background): “Recent work has applied neural networks to learn subgraph similarity for graph kernels [19].”