AI Glossary
The complete dictionary of Artificial Intelligence
Full conditional distribution
Probability distribution of a variable given all other variables in a multivariate model.
Gibbs block
Variant of Gibbs sampling where groups of variables are sampled jointly rather than individually.
Chain convergence
State where the Markov chain reaches its stationary distribution and samples reflect the target distribution.
Chain mixing
Quality of the chain to efficiently explore the entire state space without getting trapped in local regions.
Gelman-Rubin criterion
Convergence diagnostic comparing between-chain and within-chain variance across multiple MCMC chains.
Marginalization
Process of integrating out certain variables to obtain the marginal distribution of the remaining variables.
Systematic sampler
Variant of Gibbs where variables are updated in a predetermined order at each iteration.
Random sampler
Variant of Gibbs where the order of variable updates is chosen randomly at each iteration.
Conjugaison
Propriété où le prior et le posterior appartiennent à la même famille de distributions, simplifiant l'échantillonnage conditionnel.
Taille effective
Mesure du nombre d'échantillons indépendants équivalents dans une chaîne autocorrélée.