Bregman divergence based em algorithm and its application to classical and quantum rate distortion theory.
Masahito HayashiPublished in: CoRR (2022)
Keyphrases
- rate distortion theory
- bregman divergences
- information theory
- cost sensitive
- information theoretic
- rate distortion
- maximum entropy
- learning theory
- exponential family
- mahalanobis distance
- spatially adaptive
- expectation maximization
- nonnegative matrix factorization
- loss function
- nearest neighbor
- kl divergence
- special case
- mixture model
- probabilistic model
- em algorithm
- mutual information
- data dependent
- bayesian framework
- naive bayes
- gaussian distribution
- generative model
- maximum likelihood
- theoretical guarantees
- denoising
- image segmentation