Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications.
Kamilia MullakaevaLuca CosmoAnees KaziSeyed-Ahmad AhmadiNassir NavabMichael M. BronsteinPublished in: CoRR (2022)
Keyphrases
- graph representation
- graph theory
- graph structure
- directed graph
- weighted graph
- graph construction
- graph theoretic
- graph matching
- graph partitioning
- structural learning
- bipartite graph
- graph classification
- graph clustering
- spanning tree
- labeled graphs
- series parallel
- graph search
- graph theoretical
- graph databases
- graph model
- graph mining
- graph properties
- web graph
- minimum spanning tree
- spectral embedding
- learning algorithm
- graph structures
- fully connected
- adjacency matrix
- subgraph isomorphism
- graph isomorphism
- learning process
- random walk
- random graphs
- dense subgraphs
- evolving graphs
- latent variable models
- planar graphs
- undirected graph
- euclidean space
- latent variables
- structured data
- graph layout
- structural pattern recognition
- maximum common subgraph