Supervised Community Detection with Line Graph Neural Networks

By Zhengdao Chen, Lisha Li, Joan Bruna. 2017

371 citations recorded by the source corpus.

How researchers see this paper

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

In Graph Learning

Citations in the corpus

Directional Message Passing for Molecular Graphs (2020, methodology): “(Undirected) edge embeddings have already been used in previous GNNs (Jørgensen et al., 2018; Chen et al., 2019).”
Predict then Propagate: Graph Neural Networks meet Personalized PageRank (2018, background): “Deep GNNs that avoid the oversmoothing issue have been proposed in recent works by combining residual (skip) connections with batch normalization (Kawamoto et al., 2018; Chen et al., 2019).”
Scaling Graph Neural Networks with Approximate PageRank (2020, background): “Graph Neural Networks (GNNs) excel on a wide variety of network mining tasks from semi-supervised node classification and link prediction [25, 32, 44, 55] to community detection and graph classification [3, 15, 22, 37].”
Generalization and Representational Limits of Graph Neural Networks (2020, methodology): “The additional dependence on branching factor d is due to processing trees, in contrast to processing sequences in RNNs. Comparison with VC-bounds for GNNs. Scarselli et al. (2018) proved that the dependence of VC-dimension in their setting for tanh and logistic sigmoid activations is fourth order w”
Dual-Primal Graph Convolutional Networks (2018, methodology): “In a concurrent work, [9] use dual graphs in GNNs for supervised community detection, which is different from the tasks considered in this paper (classification & matrix completion).”
Universal Invariant and Equivariant Graph Neural Networks (2019, background): “However, many practical applications deal with equivariance instead, such as community detection (Chen et al., 2019), recommender systems (Ying et al.”
On the equivalence between graph isomorphism testing and function approximation with GNNs (2019, methodology): “• LGNN : Line Graph Neural Networks proposed by [8].”
Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs (2022, background): “, 2022), they also change graph structures by introducing higher-order interaction terms (Chen et al., 2019), which would require non-trivial modifications to generic GNNs in order to apply them to 3D graphs.”
Distance Encoding: Design Provably More Powerful Neural Networks for Graph Representation Learning (2020, methodology): “(4)) while GDC [32] and PowerGNN [48] used variants of ζ gpr ( u | v ) .”
Machine Learning on Graphs: A Model and Comprehensive Taxonomy (2020, positive): “Graph convolution networks Kipf et al. [61] introduce Graph Convolution Networks (GCN) which further simplify ChebNets by letting K = 2, adding a weight sharing constraint for the first and second convolutions Θ`1 = −Θ`2 := Θ`, and assuming λmax(L̃) ' 2.”
Deep Graph Library: Towards Efficient and Scalable Deep Learning on Graphs (2019, methodology): “When faced with much larger graphs, many (Hamilton et al., 2017; Chen et al., 2017) propose to sample the graph to avoid full-graph propagation.”
Can graph neural networks count substructures? (2020, methodology): “We consider LRP, GraphSAGE (using full 1-hop neighborhood) [18], GIN [64], GCN [31], 2-IGN [40], PPGN [39] and spectral GNN (sGNN) [7], with GIN and GCN under the category of MPNNs.”
Representation Learning for Dynamic Graphs: A Survey (2019, background): “Chen et al. (2019c) demonstrate that these two approaches to the expressivity of GCNs are closely related.”
Global-Local GCN: Large-Scale Label Noise Cleansing for Face Recognition (2020, methodology): “Some work [13, 33, 41, 51, 53] modiﬁes it to apply GCN into different categories, such as knowledge base construction and text classiﬁcation.”