Low-Rank Softmax Can Have Unargmaxable Classes in Theory but Rarely in Practice.
Andreas GrivasNikolay BogoychevAdam LopezPublished in: ACL (1) (2022)
Keyphrases
- low rank
- matrix factorization
- missing data
- linear combination
- convex optimization
- rank minimization
- singular value decomposition
- low rank matrix
- matrix completion
- matrix decomposition
- semi supervised
- high dimensional data
- kernel matrix
- trace norm
- high order
- low rank matrices
- minimization problems
- data matrix
- neural network
- singular values
- pairwise
- image processing
- non rigid structure from motion