Seq2Tens: An Efficient Representation of Sequences by Low-Rank Tensor Projections.
Csaba TóthPatric BonnierHarald OberhauserPublished in: CoRR (2020)
Keyphrases
- low rank
- trace norm
- high order
- matrix factorization
- frobenius norm
- missing data
- tensor decomposition
- linear combination
- convex optimization
- matrix completion
- low rank matrix
- singular value decomposition
- semi supervised
- rank minimization
- kernel matrix
- high dimensional data
- matrix decomposition
- low rank matrices
- image representation
- kullback leibler divergence
- non rigid structure from motion
- image processing
- robust principal component analysis
- singular values
- diffusion tensor
- negative matrix factorization
- multi task
- higher order
- collaborative filtering