Bypassing the Ambient Dimension: Private SGD with Gradient Subspace Identification.
Yingxue ZhouSteven WuArindam BanerjeePublished in: ICLR (2021)
Keyphrases
- dimensionality reduction
- ambient intelligence
- feature space
- principal component analysis
- edge detection
- subspace clustering
- linear subspace
- high dimensional
- subspace learning
- low dimensional
- stochastic gradient descent
- data sets
- lower dimensional
- intelligent environments
- high dimensional data
- cost function
- optical flow
- lower bound
- image processing