Graph U-Nets

By Hongyang Gao, Shuiwang Ji. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019

1,309 citations recorded by the source corpus.

We consider the problem of representation learning for graph data. Given images are special cases of graphs with nodes lie on 2D lattices, graph embedding tasks have a natural correspondence with image pixel-wise prediction tasks such as segmentation. While encoder-decoder architectures like U-Nets have been successfully applied to image pixel-wise prediction tasks, similar methods are lacking for graph data. This is because pooling and up-sampling operations are not natural on graph data. To address these challenges, we propose novel graph pooling and unpooling operations. The gPool layer adaptively selects some nodes to form a smaller graph based on their scalar projection values. We further propose the gUnpool layer as the inverse operation of the gPool layer. Based on our proposed methods, we develop an encoder-decoder model, known as the graph U-Nets. Experimental results on node classification and graph classification tasks demonstrate that our methods achieve consistently better performance than previous models. Along this direction, we extend our methods by integrating attention mechanisms. Based on attention operators, we proposed attention-based pooling and unpooling layers, which can better capture graph topology information. The empirical results on graph classification tasks demonstrate the promising capability of our methods.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/f1964cc1b343f907462cd7fbfaf88c2679d61f59

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 Computer Vision and Pattern Recognition

In Statistical and Nonlinear Physics

Citations in the corpus

Transolver: A Fast Transformer Solver for PDEs on General Geometries (2024, methodology): “Graph neural networks are representative ones, which employ kernels on connected graphs for representation learning (Hamilton et al., 2017; Gao & Ji, 2019; Pfaff et al., 2021).”
Cross-Modal Hierarchical Modelling for Fine-Grained Sketch Based Image Retrieval (2020, methodology): “Recent works like SAGPool [24] and gPool [14] focussed on top-K node selection to formulate a subgraph for the next GNN layer.”
Graph Clustering with Graph Neural Networks (2020, methodology): “Top-k [20] and SAG pooling [31] learn to sparsify the graph (select top-k edges for each node) with learned weights.”
Hierarchical Generation of Molecular Graphs using Structural Motifs (2020, methodology): “In particular, Defferrard et al. (2016) utilized graph coarsening algorithms to construct multiple layers of graph hierarchy and Ying et al. (2018); Gao & Ji (2019) proposed to learn the graph hierarchy jointly with the encoding process.”
Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition (2020, methodology): “To extract features from arbitrarily structured graphs, Graph Neural Networks (GNNs) have been developed and explored extensively [5, 17, 3, 2, 10, 40, 49, 1, 7, 11, 22].”
EvolveGCN: Evolving Graph Convolutional Networks for Dynamic Graphs (2019, background): “By convention, it takes a matrix X t with many rows as input and produces a matrix Z t with only k rows (see, e.g., (Cangea et al. 2018; Gao and Ji 2019)).”
Machine Learning on Graphs: A Model and Comprehensive Taxonomy (2020, neutral): “Other methods use differentiable pooling (Ying et al., 2018b; Cangea et al., 2018; Gao and Ji, 2019; Lee et al., 2019).”
Self-Supervised Learning of Graph Neural Networks: A Unified Review (2021, background): “With the rapid development of graph neural networks (GNNs) [23], [24], [25], [26], [27], [28], [29], basic components of GNNs [30], Yaochen Xie, Zhao Xu, Jingtun Zhang, and Shuiwang Ji are with the Department of Computer Science and Engineering, Texas A&M University, College Station, TX 77843 USA.”
Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation (2022, neutral): “1), each comprising a graph attention network (GAT) [48] and a graph-pooling layer [49]:”
ASAP: Adaptive Structure Aware Pooling for Learning Hierarchical Graph Representations (2019, methodology): “…Veliˇckovi´c et al. 2017) and link prediction (Schlichtkrull et al. 2017; Vashishth et al. 2019), others focus on learning graph-level representations for tasks like graph classiﬁcation (Bruna et al. 2013; Henaff, Bruna, and LeCun 2015; Ying et al. 2018; Gao and Ji 2019; Lee, Lee, and Kang 2019).”
XGNN: Towards Model-Level Explanations of Graph Neural Networks (2020, background): “Graph Neural Networks (GNNs) have shown their effectiveness and obtained the state-of-the-art performance on different graph tasks, such as node classification [11, 37], graph classification [39, 47], and link prediction [46].”
Medical Image Segmentation Using Deep Learning: A Survey (2020, methodology): “[140] designed a new graph pooling (gPool) and graph unpooling (gUnpool) operation based on GCN and proposed an encoder-decoder model namely graph U-Net.”
Explainability in Graph Neural Networks: A Taxonomic Survey (2020, background): “Several graphrelated tasks are widely studied, such as node classification [30], [31], [32], graph classification [5], [33], link predictions [34], [35], [36].”
Towards Deeper Graph Neural Networks (2020, background): “Great successes have been achieved for many applications, such as node classification [7, 9, 11, 12, 21, 30, 31, 33], graph classification [6, 8, 14, 18, 32, 34, 35, 38] and link prediction [1, 36, 37].”
Discovering Invariant Rationales for Graph Neural Networks (2022, methodology): “Among them, graph attention (Veliˇckovi´c et al., 2018) and pooling (Lee et al., 2019; Knyazev et al., 2019; Gao & Ji, 2019; Ranjan et al., 2020) operators prevail, which work as a computational block of a GNN to generate soft or hard masks on the input graph.”
AM-GCN: Adaptive Multi-channel Graph Convolutional Networks (2020, background): “l Networks (GCNs), a class of neural networks designed to learn graph data, have shown great popularity in tackling graph analytics problems, such as node classification [1, 31], graph classification [7, 37], link prediction [13, 36] and recommendation [6, 34]. The typical GCN [14] and its variants ”
Representing Long-Range Context for Graph Neural Networks with Global Attention (2022, methodology): “[37, 20, 14, 16, 1, etc.] and non-learned pooling methods based on classic graph coarsening schemes [10, 9, etc.].”