Random projection initialization for deep neural networks.
Piotr Iwo WójcikMarcin KurdzielPublished in: ESANN (2017)
Keyphrases
- random projections
- neural network
- dimensionality reduction
- compressed sensing
- pattern recognition
- dimension reduction
- random sampling
- image reconstruction
- hash functions
- sparse representation
- original data
- lossy image compression
- low dimensional
- principal component analysis
- high dimensional
- random samples
- data sets
- high dimensionality
- nearest neighbor
- machine learning
- document clustering
- image data