An Asymptotically Optimal Approximation of the Conditional Mean Channel Estimator Based on Gaussian Mixture Models.
Michael KollerBenedikt FeslNurettin TuranWolfgang UtschickPublished in: ICASSP (2022)
Keyphrases
- asymptotically optimal
- gaussian mixture model
- maximum likelihood
- mixture model
- asymptotic optimality
- em algorithm
- speaker recognition
- feature vectors
- heavy traffic
- arrival rate
- expectation maximization
- probability density function
- speaker identification
- variational bayes
- call center
- queueing networks
- density estimation
- generative model
- feature space
- neural network
- maximum likelihood criterion
- prior knowledge
- image segmentation