Dynamic Programming in Rank Space: Scaling Structured Inference with Low-Rank HMMs and PCFGs.
Songlin YangWei LiuKewei TuPublished in: NAACL-HLT (2022)
Keyphrases
- low rank
- dynamic programming
- frobenius norm
- missing data
- convex optimization
- linear combination
- nuclear norm
- matrix factorization
- hidden markov models
- matrix completion
- low rank matrix
- singular values
- low rank approximation
- singular value decomposition
- rank minimization
- kernel matrix
- high dimensional data
- matrix decomposition
- semi supervised
- low rank matrices
- high order
- data sets
- low rank and sparse
- inference process
- low dimensional
- linear programming
- least squares
- missing values
- minimization problems
- trace norm
- higher order
- active learning
- rank constraint
- similarity measure
- feature extraction
- face recognition