Corpus ID: 52895589

How Powerful are Graph Neural Networks?

@article{Xu2019HowPA,
  title={How Powerful are Graph Neural Networks?},
  author={Keyulu Xu and Weihua Hu and Jure Leskovec and Stefanie Jegelka},
  journal={ArXiv},
  year={2019},
  volume={abs/1810.00826}
}
Graph Neural Networks (GNNs) are an effective framework for representation learning of graphs. [...] Key Result We empirically validate our theoretical findings on a number of graph classification benchmarks, and demonstrate that our model achieves state-of-the-art performance.Expand
A Hierarchy of Graph Neural Networks Based on Learnable Local Features
TLDR
This work proposes a hierarchy of GNNs based on their aggregation regions, and derives theoretical results about the discriminative power and feature representation capabilities of each class, and shows how this framework can be utilized to systematically construct arbitrarily powerful GNN's. Expand
THE SURPRISING POWER OF GRAPH NEURAL NET-
  • 2020
Graph neural networks (GNNs) are effective models for representation learning on graph-structured data. However, standard GNNs are limited in their expressive power, as they cannot distinguish graphsExpand
k-hop Graph Neural Networks
TLDR
K-hop GNNs is proposed, which updates a node's representation by aggregating information not only from its direct neighbors, but from its k-hop neighborhood, and it is shown that the proposed architecture can identify fundamental graph properties. Expand
Graphs, Convolutions, and Neural Networks: From Graph Filters to Graph Neural Networks
TLDR
The role of graph convolutional filters in GNNs is discussed and it is shown that any architecture built with such filters has the fundamental properties of permutation equivariance and stability to changes in the topology. Expand
TOWARDS POWERFUL GRAPH NEURAL NETWORKS: DIVERSITY MATTERS
  • 2020
Graph neural networks (GNNs) offer us an effective framework for graph representation learning via layer-wise neighborhood aggregation. Their success is attributed to their expressive power atExpand
A graph similarity for deep learning
TLDR
A simple and fast GNN model based on transform-sum-cat, which obtains, in comparison with widely used GNN models, a higher accuracy in node classification, a lower absolute error in graph regression, and greater stability in adversarial training of graph generation. Expand
Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification
TLDR
This work proposes Graph Feature Network (GFN), a simple lightweight neural net defined on a set of graph augmented features, and proves that GFN can be derived by linearizing graph filtering part of GNNs, and leverages it to test the importance of the two parts separately. Expand
Self-supervised Hierarchical Graph Neural Network for Graph Representation
TLDR
This work proposes an unsupervised hierarchical neural network, referred as GraPHmax, for obtaining graph level representation and proposes the concept of periphery representation and shows its effectiveness to obtain discriminative features of an input graph. Expand
GNNExplainer: Generating Explanations for Graph Neural Networks
TLDR
GnExplainer is proposed, the first general, model-agnostic approach for providing interpretable explanations for predictions of any GNN-based model on any graph-based machine learning task. Expand
Towards Expressive Graph Representation
TLDR
This work proposes expressive GNN that aggregates the neighborhood of each node with a continuous injective set function, so that a GNN layer maps similar nodes with similar neighborhoods to similarembeddings, different nodes to different embeddings and the equivalent nodes or isomorphic graphs to the same embedDings. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 45 REFERENCES
Hierarchical Graph Representation Learning with Differentiable Pooling
TLDR
DiffPool is proposed, a differentiable graph pooling module that can generate hierarchical representations of graphs and can be combined with various graph neural network architectures in an end-to-end fashion. Expand
Representation Learning on Graphs with Jumping Knowledge Networks
TLDR
This work explores an architecture -- jumping knowledge (JK) networks -- that flexibly leverages, for each node, different neighborhood ranges to enable better structure-aware representation in graphs. Expand
An End-to-End Deep Learning Architecture for Graph Classification
TLDR
This paper designs a localized graph convolution model and shows its connection with two graph kernels, and designs a novel SortPooling layer which sorts graph vertices in a consistent order so that traditional neural networks can be trained on the graphs. Expand
Representation Learning on Graphs: Methods and Applications
TLDR
A conceptual review of key advancements in this area of representation learning on graphs, including matrix factorization-based methods, random-walk based algorithms, and graph neural networks are provided. Expand
Computational Capabilities of Graph Neural Networks
TLDR
The functions that can be approximated by GNNs, in probability, up to any prescribed degree of precision are described, and includes most of the practically useful functions on graphs. Expand
Learning Convolutional Neural Networks for Graphs
TLDR
This work proposes a framework for learning convolutional neural networks for arbitrary graphs that operate on locally connected regions of the input and demonstrates that the learned feature representations are competitive with state of the art graph kernels and that their computation is highly efficient. Expand
Anonymous Walk Embeddings
TLDR
This work coherently proposes an approach for embedding entire graphs and shows that the authors' feature representations with SVM classifier increase classification accuracy of CNN algorithms and traditional graph kernels. Expand
The Graph Neural Network Model
TLDR
A new neural network model, called graph neural network (GNN) model, that extends existing neural network methods for processing the data represented in graph domains, and implements a function tau(G,n) isin IRm that maps a graph G and one of its nodes n into an m-dimensional Euclidean space. Expand
Inductive Representation Learning on Large Graphs
TLDR
GraphSAGE is presented, a general, inductive framework that leverages node feature information (e.g., text attributes) to efficiently generate node embeddings for previously unseen data and outperforms strong baselines on three inductive node-classification benchmarks. Expand
Gated Graph Sequence Neural Networks
TLDR
This work studies feature learning techniques for graph-structured inputs and achieves state-of-the-art performance on a problem from program verification, in which subgraphs need to be matched to abstract data structures. Expand
...
1
2
3
4
5
...