Stack More Layers Differently: High-Rank Training Through Low-Rank Updates.
Vladislav LialinNamrata ShivagundeSherin MuckatiraAnna RumshiskyPublished in: CoRR (2023)
Keyphrases
- low rank
- frobenius norm
- missing data
- matrix factorization
- linear combination
- matrix completion
- nuclear norm
- singular values
- convex optimization
- singular value decomposition
- semi supervised
- matrix decomposition
- low rank approximation
- low rank matrix
- high order
- high dimensional data
- kernel matrix
- rank minimization
- training set
- trace norm
- low rank matrices
- robust principal component analysis
- data matrix
- minimization problems
- norm minimization
- collaborative filtering
- active learning
- feature selection