Perturbing Eigenvalues with Residual Learning in Graph Convolutional Neural Networks

Keyphrases

• Eigenvalues 100%
• Residual Learning 100%
• Graph Convolutional Neural Network 100%
• Graph Filters 75%
• Performance Improvement 50%
• Large-scale Networks 25%
• Easy-to-implement 25%
• Network Data 25%
• Natural Sciences 25%
• Social Sciences 25%
• Science Application 25%
• Learning Techniques 25%
• Structured Data 25%
• Spectral Domain 25%
• Eigendecomposition 25%
• Neural Unit 25%
• Aggregation Effect 25%
• Maximal Performance 25%
• Filter Matrix 25%
• Graph Convolutional Network 25%
• Process Graph 25%
• Convolutional Filter 25%
• Graph Signal 25%
• Spectral Learning 25%
• Graph-based Semi-supervised Learning 25%
• Standard Graphs 25%
• Neural Net Architecture 25%

Computer Science

• Eigenvalue 100%
• Graph Convolutional Network 100%
• Performance Improvement 50%
• Computer Network 25%
• Learning Technique 25%
• Semisupervised Learning 25%
• Structured Data 25%
• Spectral Domain 25%
• Convolutional Filter 25%

Engineering

• Eigenvalue 100%
• Convolutional Neural Network 100%
• Performance Improvement 50%
• Learning Technique 25%
• Scale Network 25%
• Spectral Domain 25%
• Neural Unit 25%