Feature selection with gradient descent on two-layer networks in low-rotation regimes.
Matus TelgarskyPublished in: CoRR (2022)
Keyphrases
- feature selection
- cost function
- feature extraction
- high dimension
- mutual information
- social networks
- network design
- rotation invariant
- complex networks
- dimensionality reduction
- method for feature selection
- neural network
- small sample
- feature selection algorithms
- unsupervised learning
- feature set
- classification accuracy
- similarity measure
- objective function
- artificial neural networks
- high dimensional
- feature space
- machine learning
- unsupervised feature selection
- bayesian networks
- selected features
- invariant features
- end to end
- loss function
- text categorization
- model selection
- support vector machine