Implicit Bias in Leaky ReLU Networks Trained on High-Dimensional Data.
Spencer FreiGal VardiPeter L. BartlettNathan SrebroWei HuPublished in: ICLR (2023)
Keyphrases
- high dimensional data
- dimensionality reduction
- nearest neighbor
- high dimensionality
- low dimensional
- high dimensional
- data sets
- subspace clustering
- high dimensions
- similarity search
- data points
- original data
- dimension reduction
- manifold learning
- high dimensional spaces
- input space
- high dimensional datasets
- data analysis
- linear discriminant analysis
- sparse representation
- data distribution
- input data
- small sample size
- gene expression data
- lower dimensional
- variable selection
- nonlinear dimensionality reduction
- training set
- clustering high dimensional data
- dimensional data
- principal component analysis
- feature extraction
- locally linear embedding
- text data
- data streams
- high dimensional data sets
- learning algorithm