Transductive nonnegative matrix factorization for semi-supervised high-performance speech separation.
Naiyang GuanLong LanDacheng TaoZhigang LuoXuejun YangPublished in: ICASSP (2014)
Keyphrases
- nonnegative matrix factorization
- semi supervised
- semi supervised clustering
- semi supervised learning
- manifold regularization
- labeled data
- probabilistic latent semantic indexing
- unlabeled data
- semi supervised classification
- pairwise
- data representation
- supervised learning
- active learning
- negative matrix factorization
- unsupervised learning
- pairwise constraints
- spectral clustering
- nonnegative matrix
- matrix factorization
- least squares
- document clustering
- metric learning
- low rank
- principal component analysis
- labeled and unlabeled data
- data sets
- collaborative filtering
- probabilistic model
- objective function
- sparsity constraints
- similarity measure
- learning algorithm