Dynamical mean-field theory for stochastic gradient descent in Gaussian mixture classification.
Francesca MignaccoFlorent KrzakalaPierfrancesco UrbaniLenka ZdeborováPublished in: NeurIPS (2020)
Keyphrases
- gaussian mixture
- stochastic gradient descent
- expectation maximization
- feature selection
- gaussian mixture model
- feature vectors
- support vector
- feature space
- density estimation
- supervised learning
- image classification
- loss function
- closed form
- machine learning
- feature extraction
- matrix factorization
- unsupervised learning
- text classification
- support vector machine
- model selection
- mixture model
- em algorithm
- multiscale
- machine learning methods
- markov random field
- least squares
- probability distribution