Phase diagram of Stochastic Gradient Descent in high-dimensional two-layer neural networks.
Rodrigo VeigaLudovic StephanBruno LoureiroFlorent KrzakalaLenka ZdeborováPublished in: NeurIPS (2022)
Keyphrases
- stochastic gradient descent
- high dimensional
- neural network
- matrix factorization
- loss function
- least squares
- step size
- dimensionality reduction
- random forests
- low dimensional
- multiple kernel learning
- regularization parameter
- data points
- support vector machine
- collaborative filtering
- importance sampling
- online algorithms
- nearest neighbor
- genetic algorithm
- high dimensional data
- input space
- convergence rate
- training set
- feature space
- random forest
- parameter space
- weight vector
- decision trees
- image processing
- machine learning