Does SGD really happen in tiny subspaces?
Minhak SongKwangjun AhnChulhee YunPublished in: CoRR (2024)
Keyphrases
- high dimensional data
- stochastic gradient descent
- low dimensional
- feature space
- high dimensional
- principal component analysis
- grassmann manifold
- canonical correlations
- mutual subspace method
- neural network
- high dimensional feature spaces
- stochastic gradient
- hilbert space
- lower dimensional
- linear subspace
- clustering method
- linear combination
- dimensionality reduction