When Work Matters: Transforming Classical Network Structures to Graph CNN

Numerous pattern recognition applications can be formed as learning from graph-structured data, including social network, protein-interaction network, the world wide web data, knowledge graph, etc. While convolutional neural network (CNN) facilitates great advances in gridded image/video understanding tasks, very limited attention has been devoted to transform these successful network structures (including Inception net, Residual net, Dense net, etc.).. (read more)

PDF Abstract
No code implementations yet. Submit your code now

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Graph Classification COLLAB G_DenseNet Accuracy 83.16% # 3
Graph Classification ENZYMES G_Inception Accuracy 67.50% # 6
Graph Classification IMDb-B G_ResNet Accuracy 79.90% # 2
Graph Classification IMDb-M G_ResNet Accuracy 54.53% # 3
Graph Classification MUTAG G_Inception Accuracy 95.00% # 1
Graph Classification NCI109 G_DenseNet Accuracy 80.66 # 10
Graph Classification PTC G_DenseNet Accuracy 73.24% # 5

Methods used in the Paper


METHOD TYPE
Average Pooling
Pooling Operations
Concatenated Skip Connection
Skip Connections
Dense Block
Image Model Blocks
Dropout
Regularization
Dense Connections
Feedforward Networks
Softmax
Output Functions
DenseNet
Convolutional Neural Networks
Residual Connection
Skip Connections
ReLU
Activation Functions
1x1 Convolution
Convolutions
Batch Normalization
Normalization
Bottleneck Residual Block
Skip Connection Blocks
Global Average Pooling
Pooling Operations
Residual Block
Skip Connection Blocks
Kaiming Initialization
Initialization
Max Pooling
Pooling Operations
Convolution
Convolutions
ResNet
Convolutional Neural Networks