Matrix Pontryagin principle approach to controllability metrics maximization under sparsity constraints.
Tomofumi OhtsukaTakuya IkedaKenji KashimaPublished in: CoRR (2022)
Keyphrases
- sparsity constraints
- low rank
- factor matrices
- matrix factorization
- nonnegative matrix factorization
- singular value decomposition
- sparse representation
- negative matrix factorization
- dictionary learning
- linear combination
- sparse coding
- objective function
- convex optimization
- high order
- global constraints
- kernel matrix
- data matrix
- computer vision
- higher order
- semi supervised
- high dimensional data
- pattern recognition
- multiscale