Extensive experiments show that graph classification accuracy with RGM feature maps is better than or competitive with many powerful graph kernels, unsupervised ...
Distribution of Node Embeddings as Multiresolution Features for Graphs
ieeexplore.ieee.org › document
To address such challenges, we propose Randomized Grid Mapping (RGM), a fast-to-compute feature map that represents a graph via the distribution of its node ...
Distribution of Node Embeddings as Multiresolution Features for Graphs
www.computer.org › csdl › icdm
1: Overview of our framework. · 2: Multiresolution feature maps for graphs. · 3: Upper left quadrant is best: Accuracy vs runtime for RGM and its closest ...
Abstract—Graph classification is an important problem in many fields, from bioinformatics and neuroscience to computer vision and social network analysis.
For comparing entire graphs, aggregating node embeddings (as we do) is competitive to deep neural networks, graph kernels, and feature construction [9] .
Randomized Grid Mapping is proposed, a fast-to-compute feature map that represents a graph via the distribution of its node embeddings in feature space that ...
Heimann, Mark, Safavi, Tara, and Koutra, Danai. "Distribution of Node Embeddings as Multiresolution Features for Graphs". IEEE International Conference on Data ...
Bibliographic details on Distribution of Node Embeddings as Multiresolution Features for Graphs.
7 days ago · On the other hand, while the same input network can be represented at multiple levels of resolution by coarse-graining the constituent nodes ...
RGM characterizes a graph by the distribution of its node's latent features or embeddings in vector space. The resulting feature maps may be used to perform ...