k-hop graph neural networks

Nikolentzos, Giannis; Dasoulas, George; Vazirgiannis, Michalis

doi:10.1016/j.neunet.2020.07.008

Cited by 73 publications

(73 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(2) The state-of-the-art GNNs: Graph convolution network (GCN) [37], Deep Graph CNN (DGCNN) [38], Graph Isomorphism Network (GIN) [3], Random Walk Graph Neural Network (RW-GNN) [39], Graph Attention Network (GAT) [2], Motif based Attentional Graph Convolutional Neural Network (MA-GCNN) [40].…”

Section: B Baselines For Comparisonmentioning

confidence: 99%

“…For a fair comparison, all of the methods are run on a system with an Intel Xeon CPU E3-1270 v5 processor. Due to RW-GNN cannot accept NCI1 as input [39], the running time of RW-GNN on NCI1 dataset is vacant. Obviously, TL-GNN requires more time than GIN because of the extra computation of subgraph-level and attention mechanisms.…”

Section: Comparison Of Computational Complexitymentioning

confidence: 99%

See 1 more Smart Citation

Two-Level Graph Neural Network

Sun

Zhang

et al. 2024

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) are recently proposed neural network structures for the processing of graphstructured data. Due to their employed neighbor aggregation strategy, existing GNNs focus on capturing node-level information and neglect high-level information. Existing GNNs therefore suffer from representational limitations caused by the Local Permutation Invariance (LPI) problem. To overcome these limitations and enrich the features captured by GNNs, we propose a novel GNN framework, referred to as the Two-level GNN (TL-GNN). This merges subgraph-level information with nodelevel information. Moreover, we provide a mathematical analysis of the LPI problem which demonstrates that subgraph-level information is beneficial to overcoming the problems associated with LPI. A subgraph counting method based on the dynamic programming algorithm is also proposed, and this has time complexity is O(n 3 ), n is the number of nodes of a graph. Experiments show that TL-GNN outperforms existing GNNs and achieves state-of-the-art performance.

show abstract

Section: B Baselines For Comparisonmentioning

confidence: 99%

Section: Comparison Of Computational Complexitymentioning

confidence: 99%

Two-Level Graph Neural Network

Sun

Zhang

et al. 2024

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…While GCKN utilizes the local walk and path only starting from the central node, our model considers any walks (up to a maximal length) within the subgraph around the central node, and can thus explore more topological structures. Another recent work by Nikolentzos and Vazirgiannis (2020) focused on improving model transparency by calculating the graph kernels between trainable hidden graphs and the entire graph. However, the method only supports a single-layer model and lacks theoretical interpretation.…”

Section: Combination Of Graph Kernel and Gnnsmentioning

confidence: 99%

KerGNNs: Interpretable Graph Neural Networks with Graph Kernels

Feng¹,

You²,

Wang³

2022

Preprint

View full text Add to dashboard Cite

Graph kernels are historically the most widely-used technique for graph classification tasks. However, these methods suffer from limited performance because of the hand-crafted combinatorial features of graphs. In recent years, graph neural networks (GNNs) have become the state-of-the-art method in downstream graph-related tasks due to their superior performance. Most GNNs are based on Message Passing Neural Network (MPNN) frameworks. However, recent studies show that MPNNs can not exceed the power of the Weisfeiler-Lehman (WL) algorithm in graph isomorphism test. To address the limitations of existing graph kernel and GNN methods, in this paper, we propose a novel GNN framework, termed Kernel Graph Neural Networks (KerGNNs), which integrates graph kernels into the message passing process of GNNs. Inspired by convolution filters in convolutional neural networks (CNNs), KerGNNs adopt trainable hidden graphs as graph filters which are combined with subgraphs to update node embeddings using graph kernels. In addition, we show that MPNNs can be viewed as special cases of KerGNNs. We apply KerGNNs to multiple graph-related tasks and use cross-validation to make fair comparisons with benchmarks. We show that our method achieves competitive performance compared with existing state-of-the-art methods, demonstrating the potential to increase the representation ability of GNNs. We also show that the trained graph filters in KerGNNs can reveal the local graph structures of the dataset, which significantly improves the model interpretability compared with conventional GNN models.In recent years, the machine learning research community has devoted substantial energy to applying graph neural networks (GNNs) to numerous downstream graph-related tasks (

show abstract

“…Du et al (2019) followed the opposite direction and proposed a new graph kernel which corresponds to infinitely wide multi-layer GNNs trained by gradient descent, while Al-Rfou et al (2019) proposed an unsupervised method for learning graph representations by comparing the input graphs against a set of source graphs. Finally, Nikolentzos and Vazirgiannis (2020) proposed a neural network model whose first layer consists of a number of latent graphs which are compared against the input graphs using a random walk kernel. The emerging kernel values are fed into a fully-connected neural network which acts as the classifier or regressor.…”

Section: Other Modelsmentioning

confidence: 99%

Graph Kernels: A Survey

Nikolentzos

Siglidis²,

Vazirgiannis

2021

jair

Self Cite

View full text Add to dashboard Cite

Graph kernels have attracted a lot of attention during the last decade, and have evolved into a rapidly developing branch of learning on structured data. During the past 20 years, the considerable research activity that occurred in the field resulted in the development of dozens of graph kernels, each focusing on specific structural properties of graphs. Graph kernels have proven successful in a wide range of domains, ranging from social networks to bioinformatics. The goal of this survey is to provide a unifying view of the literature on graph kernels. In particular, we present a comprehensive overview of a wide range of graph kernels. Furthermore, we perform an experimental evaluation of several of those kernels on publicly available datasets, and provide a comparative study. Finally, we discuss key applications of graph kernels, and outline some challenges that remain to be addressed.

show abstract

k-hop graph neural networks

Cited by 73 publications

References 17 publications

Two-Level Graph Neural Network

Two-Level Graph Neural Network

KerGNNs: Interpretable Graph Neural Networks with Graph Kernels

Graph Kernels: A Survey

Contact Info

Product

Resources

About