Homeomorphic-Invariance of EM: Non-Asymptotic Convergence in KL Divergence for Exponential Families via Mirror Descent.
Frederik KunstnerRaunak KumarMark SchmidtPublished in: CoRR (2020)
Keyphrases
- exponential family
- kl divergence
- maximum likelihood
- mixture model
- density estimation
- expectation maximization
- em algorithm
- log likelihood
- graphical models
- missing values
- closed form
- statistical models
- gaussian mixture model
- probabilistic model
- generative model
- parameter estimation
- hidden variables
- gaussian mixture
- order statistics
- maximum a posteriori
- probability density function
- variational methods
- gaussian distribution
- machine learning
- bayesian framework
- unsupervised learning