OLÉ: Orthogonal Low-Rank Embedding - A Plug and Play Geometric Loss for Deep Learning.
José LezamaQiang QiuPablo MuséGuillermo SapiroPublished in: CVPR (2018)
Keyphrases
- low rank
- deep learning
- high order
- matrix factorization
- linear combination
- missing data
- convex optimization
- unsupervised learning
- high dimensional data
- matrix completion
- semi supervised
- rank minimization
- machine learning
- mental models
- singular value decomposition
- weakly supervised
- singular values
- higher order
- nearest neighbor
- data analysis
- dimensionality reduction
- collaborative filtering
- high dimensional
- image processing