High-Dimensional Neural Feature Using Rectified Linear Unit And Random Matrix Instance.
Alireza M. JavidArun VenkitaramanMikael SkoglundSaikat ChatterjeePublished in: ICASSP (2020)
Keyphrases
- high dimensional
- low dimensional
- neural network
- network architecture
- dimensionality reduction
- multi dimensional
- image features
- pseudo inverse
- square root
- eigenvalue decomposition
- similarity search
- semidefinite
- nearest neighbor
- bio inspired
- stereo images
- weight vector
- dimension reduction
- singular value decomposition
- covariance matrix
- neural model
- sparse data
- linear systems
- linearly independent
- linear model
- low rank
- associative memory
- metric space
- parameter space
- microarray data
- high dimensionality
- natural images
- sufficient conditions
- feature set