Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification.
Francesca MignaccoFlorent KrzakalaPierfrancesco UrbaniLenka ZdeborováPublished in: CoRR (2020)
Keyphrases
- gaussian mixture
- stochastic gradient descent
- feature selection
- machine learning methods
- machine learning
- support vector machine
- image classification
- gaussian mixture model
- feature space
- decision trees
- density estimation
- text classification
- support vector
- em algorithm
- model selection
- feature extraction
- probability density function
- matrix factorization
- step size
- unsupervised learning
- machine learning algorithms
- cross validation
- closed form
- markov random field
- least squares
- probability density
- multiple kernel learning