YZ Sözlüğü
Yapay Zekanın tam sözlüğü
Expectation-Maximization (EM) Algorithm
An iterative method to find maximum likelihood estimates of parameters in probabilistic models with latent variables, alternating between an expectation (E) step and a maximization (M) step.
Conjugate Prior
The choice of a prior distribution which, when combined with the likelihood function, yields a posterior distribution belonging to the same family of distributions, greatly simplifying analytical calculations.
Markov Chain Monte Carlo (MCMC) Method
A class of algorithms for sampling from probability distributions by constructing a Markov chain whose stationary distribution is the target distribution, allowing for the approximation of Bayesian integrals.
Mean Field Approximation
A family of variational approximations that assumes the posterior distribution factorizes into a product of simpler distributions for each variable, simplifying optimization at the cost of losing correlations.
Epistemic Inference
A branch of inference that focuses on quantifying uncertainty due to a lack of knowledge, often modeled by probability distributions over model parameters in a Bayesian framework.