• Corpus ID: 208006339

Hierarchical Graph Pooling with Structure Learning

  title={Hierarchical Graph Pooling with Structure Learning},
  author={Zhen Zhang and Jiajun Bu and Martin Ester and Jianfeng Zhang and Chengwei Yao and Zhi Yu and Can Wang},
Graph Neural Networks (GNNs), which generalize deep neural networks to graph-structured data, have drawn considerable attention and achieved state-of-the-art performance in numerous graph related tasks. However, existing GNN models mainly focus on designing graph convolution operations. The graph pooling (or downsampling) operations, that play an important role in learning hierarchical representations, are usually overlooked. In this paper, we propose a novel graph pooling operator, called… 

Figures and Tables from this paper

Hierarchical Multi-View Graph Pooling With Structure Learning

A novel multi-view graph pooling operator dubbed as MVPool, which ranks nodes across different views with different contextual graph information is proposed, which performs hierarchical representation learning for both node and graph level classification as well as clustering tasks.

Graph Multihead Attention Pooling with Self-Supervised Learning

This paper proposes a hierarchical graph pooling method based on the multihead attention mechanism, namely GMAPS, which compresses both node features and graph structure into the coarsened graph.

DMSPool: Dual Multi-Scale Pooling for Graph Representation Learning

Dual Multi-Scale Pooling (DMSPool) is proposed, which uses multiple architectures concurrently to integrate graph convolution and pooling modules in an end-to-end fashion and achieves superior graph classification performance over the state-of-the-art graph representation learning methods.

Hierarchical Graph Pooling With Self-Adaptive Cluster Aggregation

Experimental results show that combining the existing GNN architecture with HGP-SACA can achieve state-of-the-art results on multiple graph classification benchmarks, which proves the effectiveness of the proposed model.

Multi-Channel Pooling Graph Neural Networks

This work proposes a Multi-channel Graph Pooling method, which captures the local structure, the global structure and node features simultaneously in graph pooling, and presents the superior performance of MuchPool.

Deep Graph Structure Learning for Robust Representations: A Survey

A general paradigm of Graph Structure Learning is formulated, and state-ofthe-art methods classified by how they model graph structures are reviewed, followed by applications that incorporate the idea of GSL in other graph tasks.

Supervised Contrastive Learning with Structure Inference for Graph Classification

This paper proposes a novel graph neural network based on supervised contrastive learning with structure inference for graph classification that can discover additional connections to enhance the existing edge set and resorts to a structure inference stage based on diffusion cascades to recover possible connections with high node similarities.

Distribution Knowledge Embedding for Graph Pooling

It is argued what is crucial to graph-level downstream tasks includes not only the topological structure but also the distribution from which nodes are sampled, and a new plug-and-play pooling module, termed as Distribution Knowledge Embedding (DKEPool), is proposed.

Edge but not Least: Cross-View Graph Pooling

This paper presents a cross-view graph pooling method (Co-Pooling) that explicitly exploits crucial graph substructures for learning graph representations and can achieve superior performance over state-of-the-art pooling methods on graph classification and regression tasks.