hhh920406/DeepGCN

Deep Graph Convolutional Networks

GCN

Node Classification

• Semi-Supervised Classification with Graph Convolutional Networks (GCN)
• Inductive Representation Learning on Large Graphs (GraphSAGE)
• Graph Attention Networks (GAT)
• Representation Learning on Graphs with Jumping Knowledge Networks
• Simplifying Graph Convolutional Networks (SGC)
  • Code (PyTorch)
• Predict then Propagate: Graph Neural Networks meet Personalized PageRank (APPNP)
• Deep Graph Infomax
• MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing
• Position-aware Graph Neural Networks
• Disentangled Graph Convolutional Networks
• A Representation Learning Framework for Property Graphs
• Exploiting Edge Features for Graph Neural Networks
• Power up! Robust Graph Convolutional Network against Evasion Attacks based on Graph Powering
  • Read carefully. Re-examine Laplacian operator - find some basic flaws in the spatial and spectral domains. Propose a new operator.

Graph Classification

• How Powerful are Graph Neural Networks? (GIN)
• An End-to-End Deep Learning Architecture for Graph Classification (DGCNN)
  • Code (Torch), Code (PyTorch)
• Graph Capsule Convolutional Neural Networks
• Capsule Graph Neural Network
  • Code (PyTorch), Code (TF)
• Discriminative structural graph classification
  • Discrimination capacity of aggregation functions
• Relational Pooling for Graph Representations
  • Read carefully. Use node and edge features.
• Optimal Transport for structured data with application on graphs
  • See Wasserstein Weisfeiler-Lehman Graph Kernels

Dynamic

• Dynamic Graph Message Passing Networks

Graph Learning

• Learning Discrete Structures for Graph Neural Networks
  • Sample graphs, learn distribution of each graph, and node classification. Transductive.
• Exploring Graph Learning for Semi-Supervised Classification Beyond Euclidean Data
  • Learn adj matrix.
• Graph Learning Networks (New problem)
  • Build an explicit graph and learn the graph.
• Graph Matching Networks for Learning the Similarity of Graph Structured Objects
  • Graph similarity.
• Learning to Route in Similarity Graphs
  • Routing problem in graphs. Neareast neighbor search.
• Large Scale Graph Learning From Smooth Signals
  • Graph construction by approximate neareat neighbor techniques. Make it efficient.
• Spectral Inference Networks: Unifying Deep and Spectral Learning

Unsupervised Learning

• Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
• Pre-training Graph Neural Networks
• Disentangled Graph Convolutional Networks
• Deep Graph Infomax
• UNSUPERVISED PRE-TRAINING OF GRAPH CONVOLUTIONAL NETWORKS
• Graphite: Iterative Generative Modeling of Graphs
  • An algorithmic framework for unsupervised learning of representations over nodes in large graphs using deep latent variable generative models.
• Unsupervised Inductive Graph-Level Representation Learning via Graph-Graph Proximity

Adversarial Attacks

• Adversarial attacks on neural networks for graph data
• Adversarial attack on graph structured data
• Adversarial attacks on graph neural networks via meta learning
• Attacking Graph Convolutional Networks via Rewiring (New problem)
  • Rewiring attack. Perturbation
• Robust Graph Convolutional Networks Against Adversarial Attacks
• Stability Properties of Graph Neural Networks
  • Permutation Equivariant. Stability properties in two perturbation models.
• Stability of Graph Scattering Transforms
• Stability and Generalization of Graph Convolutional Neural Networks

Theoretic analysis on GNN

• Approximation Ratios of Graph Neural Networks for Combinatorial Problems

Self-attention CNN

• Stand-Alone Self-Attention in Vision Models
• Graph Attention Networks (GAT)

Position-aware GCN

• Position-aware Graph Neural Networks

Not published

Node Classification

• Revisiting Graph Neural Networks: All We Have is Low-Pass Filters
• Edge Contraction Pooling for Graph Neural Networks (Node and graph classification)
• Variational Spectral Graph Convolutional Networks
  • Noisy graphs.
• Dimensional Reweighting Graph Convolutional Networks
• GraphSAINT: Graph Sampling Based Inductive Learning Method
• Spectral-based Graph Convolutional Network for Directed Graphs

Graph Classification

• Are Powerful Graph Neural Nets Necessary? A Dissection on Graph Classification
• Neighborhood Enlargement in Graph Neural Networks
• Provably Powerful Graph Networks, Accept
• Wasserstein Weisfeiler-Lehman Graph Kernels, Accept
• Graph Filtration Learning
• Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation
• Understanding the Representation Power of Graph Neural Networks in Learning Graph Topology, Accept

Generative Model

• MolecularRNN: Generating realistic molecular graphs with optimized properties
• A Two-Step Graph Convolutional Decoder for Molecule Generation
• GRAM: Scalable Generative Models for Graphs with Graph Attention Mechanism

Others

• Relational Reasoning using Prior Knowledge for Visual Captioning
• Discovering Neural Wirings
• LEARNING REPRESENTATIONS OF GRAPH DATA: A SURVEY
• Redundancy-Free Computation Graphs for Graph Neural Networks
• Representation Learning on Networks: Theories, Algorithms, and Applications, [Tutorial]
• Attributed Graph Clustering: A Deep Attentional Embedding Approach
• Hybrid Low-order and Higher-order Graph Convolutional Networks

CNN Ideas

• Convolutional Neural Networks for Sentence Classification
• A Convolutional Neural Network for Modelling Sentences
  • 참고, 참고2

Knowlede Graphs

• Learning to Exploit Long-term Relational Dependencies in Knowledge Graphs