Geodesically-convex optimization for averaging partially observed covariance matrices.
Florian YgerSylvain ChevallierQuentin BarthélemySuvrit SraPublished in: ACML (2020)
Keyphrases
- convex optimization
- covariance matrices
- partially observed
- covariance matrix
- maximum likelihood
- distance measure
- gaussian mixture model
- vector space
- gaussian distribution
- low rank
- total variation
- augmented lagrangian
- feature vectors
- gaussian mixture
- linear classifiers
- probability density function
- lie group
- learning algorithm
- em algorithm
- higher order
- support vector