AI-ordlista
Den kompletta ordlistan över AI
Metropolis-Hastings algorithm
Fundamental MCMC algorithm using a proposal distribution to generate samples from the target distribution.
Gibbs Sampling
Special case of Metropolis-Hastings where each variable is sampled conditionally on all others.
Hamiltonian MCMC
Uses Hamiltonian mechanics and gradients to efficiently explore the parameter space.
MCMC adaptatif
Algorithmes qui ajustent automatiquement leurs paramètres pendant l'exécution pour optimiser l'efficacité.
Convergence diagnostics
Statistical methods to evaluate whether MCMC chains have reached their stationary distribution.
Parallel MCMC
Techniques for running multiple chains simultaneously and combining their results efficiently.
Reversible jump MCMC
MCMC extension allowing jumps between models with different dimensions.
Slice sampling
MCMC method that samples from the distribution under a random slice of the density.
MCMC for hierarchical models
Specialized application of MCMC for inference in hierarchical Bayesian models.
High-dimensional MCMC
Techniques tailored to handle the curse of dimensionality in complex parameter spaces.
Warm-up chains
Initial adaptation phase before sample collection to reach the stationary distribution.
Variational MCMC
Hybrid between variational inference and MCMC to accelerate convergence.