Prior and posterior
Prior and posterior: Distributions representing parameter knowledge before data and updated knowledge after observations, respectively. The posterior combines prior and likelihood through Bayes’ rule. Comparing both shows how much observations changed the initial uncertainty. This perspective connects the definition with practical behaviour in artificial intelligence systems.
In italiano: Prior e PosteriorLegge II · Probabilità, softmax e modelli generativiThe posterior combines prior and likelihood through Bayes’ rule. Comparing both shows how much observations changed the initial uncertainty.