Node Similarity Preserving Graph Convolutional Networks

Jin, Wei; Derr, Tyler; Wang, Yiqi; Ma, Yao; Liu, Zitao; Tang, Jiliang

doi:10.1145/3437963.3441735

Cited by 169 publications

(99 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Non-Homophilous methods. Various GNNs have been proposed to achieve higher performance in low-homophily settings [82,44,81,17,15,73,36,35]. Geom-GCN [58] introduces a geometric aggregation scheme, MixHop [1] proposes a graph convolutional layer that mixes powers of the adjacency matrix, GPR-GNN [17] features learnable weights that can be positive and negative in feature propagation, GCNII [15] allows deep graph convolutional networks with relieved oversmoothing, which empirically performs better in non-homophilous settings, and H 2 GCN [82] shows that separation of ego and neighbor embeddings, aggregation in higher-order neighborhoods, and the combination of intermediate representations improves GNN performance in low-homophily.…”

Section: Prior Workmentioning

confidence: 99%

“…For example, malicious node detection, a key application of graph machine learning, is known to be non-homophilous in many settings [55,13,25,11]. Further, while new GNNs that work better in these non-homophilous settings have been developed [82,44,81,17,15,73,36,35,9,54], their evaluation is limited to a few graph datasets used by Pei et al [58] (collected by [61,66,48]) that have certain undesirable properties such as small size, narrow range of application areas, and high variance between different train/test splits [82]. Consequently, method scalability has not been thoroughly studied in non-homophilous graph learning.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods

Lim¹,

Höhne²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

Many widely used datasets for graph machine learning tasks have generally been homophilous, where nodes with similar labels connect to each other. Recently, new Graph Neural Networks (GNNs) have been developed that move beyond the homophily regime; however, their evaluation has often been conducted on small graphs with limited application domains. We collect and introduce diverse nonhomophilous datasets from a variety of application areas that have up to 384x more nodes and 1398x more edges than prior datasets. We further show that existing scalable graph learning and graph minibatching techniques lead to performance degradation on these non-homophilous datasets, thus highlighting the need for further work on scalable non-homophilous methods. To address these concerns, we introduce LINKX -a strong simple method that admits straightforward minibatch training and inference. Extensive experimental results with representative simple methods and GNNs across our proposed datasets show that LINKX achieves state-ofthe-art performance for learning on non-homophilous graphs. Our codes and data are available at https://github.com/CUAI/Non-Homophily-Large-Scale.

show abstract

Section: Prior Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods

Lim¹,

Höhne²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…In this paper, we focus on perturbing the structure by adding or deleting edges, and evaluating the robustness of our methods on the node classification task. Specifically, we use metattack [40] to perform non-targeted attack, and follow the same experimental setting as [15], i.e., the ratio of changed edges, from 0 to 25% with a step of 5%. We use GCN, GAT, APPNP and FAGCN as baselines and use the default hyperparameter settings in the authors' implementations.…”

Section: Adversarial Defensementioning

confidence: 99%

Graph Neural Networks with Feature and Structure Aware Random Walk

Zhuo¹,

Yu²,

Tan³

2021

Preprint

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) have received increasing attention for representation learning in various machine learning tasks. However, most existing GNNs applying neighborhood aggregation usually perform poorly on the graph with heterophily where adjacent nodes belong to different classes. In this paper, we show that in typical heterphilous graphs, the edges may be directed, and whether to treat the edges as is or simply make them undirected greatly affects the performance of the GNN models. Furthermore, due to the limitation of heterophily, it is highly beneficial for the nodes to aggregate messages from similar nodes beyond local neighborhood.These motivate us to develop a model that adaptively learns the directionality of the graph, and exploits the underlying long-distance correlations between nodes. We first generalize the graph Laplacian to digraph based on the proposed Feature-Aware PageRank algorithm, which simultaneously considers the graph directionality and long-distance feature similarity between nodes. Then digraph Laplacian defines a graph propagation matrix that leads to a model called DiglacianGCN. Based on this, we further leverage the node proximity measured by commute times between nodes, in order to preserve the nodes' long-distance correlation on the topology level. Extensive experiments on ten datasets with different levels of homophily demonstrate the effectiveness of our method over existing solutions in the task of node classification. CCS CONCEPTS• Mathematics of computing → Graph algorithms; • Computing methodologies → Learning latent representations; Neural networks.

show abstract

“…Low homophily level in some real-word graphs can largely degrade their performance . Some initial efforts (Pei et al 2020;Bo et al 2021;Jin et al 2021;) have been taken to address the problem of heterophilic graphs. For example, H2GCN ) investigated three key designs for GNNs on heterophilic graphs.…”

Section: Related Workmentioning

confidence: 99%

“…In other words, we obtain more discriminative features. Though promising, there is no existing work exploring label-wise message passing to address the challenge of heterophilic graphs (Bo et al 2021;Jin et al 2021).…”

Section: Introductionmentioning

confidence: 99%

Label-Wise Message Passing Graph Neural Network on Heterophilic Graphs

Dai¹,

Guo²,

Wang

2021

Preprint

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) have achieved remarkable performance in modeling graphs for various applications. However, most existing GNNs assume the graphs exhibit strong homophily in node labels, i.e., nodes with similar labels are connected in the graphs. They fail to generalize to heterophilic graphs where linked nodes may have dissimilar labels and attributes. Therefore, in this paper, we investigate a novel framework that performs well on graphs with either homophily or heterophily. More specifically, to address the challenge brought by the heterophily in graphs, we propose a label-wise message passing mechanism. In label-wise message-passing, neighbors with similar pseudo labels will be aggregated together, which will avoid the negative effects caused by aggregating dissimilar node representations. We further propose a bi-level optimization method to automatically select the model for graphs with homophily/heterophily. Extensive experiments demonstrate the effectiveness of our proposed framework for node classification on both homophilic and heterophilic graphs.

show abstract

Node Similarity Preserving Graph Convolutional Networks

Cited by 169 publications

References 6 publications

Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods

Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods

Graph Neural Networks with Feature and Structure Aware Random Walk

Label-Wise Message Passing Graph Neural Network on Heterophilic Graphs

Contact Info

Product

Resources

About