Mean-Field Microcanonical Gradient Descent.
Marcus HäggbomMorten KarlsmarkJoakim AndénPublished in: CoRR (2024)
Keyphrases
- cost function
- markov random field
- loss function
- objective function
- statistical mechanics
- belief networks
- em algorithm
- bayesian inference
- markov networks
- closed form
- deterministic annealing
- free energy
- learning rules
- figure ground separation
- linear complexity
- error function
- artificial neural networks
- data clustering
- belief propagation
- loopy belief propagation
- mumford shah functional
- conjugate gradient
- parameter estimation
- maximum likelihood
- coordinate ascent