High-Dimensional Private Empirical Risk Minimization by Greedy Coordinate Descent.
Paul MangoldAurélien BelletJoseph SalmonMarc TommasiPublished in: CoRR (2022)
Keyphrases
- empirical risk minimization
- high dimensional
- statistical learning theory
- uniform convergence
- low dimensional
- empirical risk
- dimensionality reduction
- rates of convergence
- similarity search
- phase transition
- vc dimension
- feature selection
- logistic regression
- support vector machine
- variable selection
- feature space
- nearest neighbor
- upper bound
- kernel learning
- decision function
- loss function
- supervised classification
- statistical learning
- input space
- metric space
- high dimensional data
- kernel function
- training samples
- model selection
- sufficient conditions
- data points
- decision trees