Unsupervised Learning of Graph Hierarchical Abstractions with Differentiable Coarsening and Optimal Transport.
Tengfei MaJie ChenPublished in: CoRR (2019)
Keyphrases
- unsupervised learning
- structured data
- dynamic programming
- supervised learning
- graph partitioning
- graph representation
- objective function
- minimum spanning tree
- semi supervised
- deep learning
- graph matching
- loss function
- pairwise
- optimal solution
- training data
- minimum cost
- graph cuts
- data clustering
- connected components
- closed form
- data sets
- dimensionality reduction
- worst case
- pattern recognition
- high level