How to iron out rough landscapes and get optimal performances: Replicated Gradient Descent and its application to tensor PCA.
Giulio BiroliChiara CammarotaFederico Ricci-TersenghiPublished in: CoRR (2019)
Keyphrases
- principal component analysis
- dimensionality reduction
- tensor analysis
- principal components analysis
- fault tolerant
- high order
- higher order
- high dimensional
- feature space
- optimal solution
- loss function
- objective function
- covariance matrix
- principal components
- dynamic programming
- worst case
- closed form
- discriminant analysis
- feature extraction
- computer vision
- principle component analysis
- learning algorithm