Module: ed.inferences

Defined in edward/inferences/__init__.py.

Classes

class BiGANInference: Adversarially Learned Inference (Dumoulin et al., 2017) or

class GANInference: Parameter estimation with GAN-style training (Goodfellow et al.,

class Gibbs: Gibbs sampling (Geman and Geman, 1984).

class HMC: Hamiltonian Monte Carlo, also known as hybrid Monte Carlo

class ImplicitKLqp: Variational inference with implicit probabilistic models

class Inference: Abstract base class for inference. All inference algorithms in

class KLpq: Variational inference with the KL divergence

class KLqp: Variational inference with the KL divergence

class Laplace: Laplace approximation (Laplace, 1774).

class MAP: Maximum a posteriori.

class MetropolisHastings: Metropolis-Hastings (Metropolis et al., 1953; Hastings, 1970).

class MonteCarlo: Abstract base class for Monte Carlo. Specific Monte Carlo methods

class ReparameterizationEntropyKLqp: Variational inference with the KL divergence

class ReparameterizationKLKLqp: Variational inference with the KL divergence

class ReparameterizationKLqp: Variational inference with the KL divergence

class SGHMC: Stochastic gradient Hamiltonian Monte Carlo (Chen et al., 2014).

class SGLD: Stochastic gradient Langevin dynamics (Welling and Teh, 2011).

class ScoreEntropyKLqp: Variational inference with the KL divergence

class ScoreKLKLqp: Variational inference with the KL divergence

class ScoreKLqp: Variational inference with the KL divergence

class VariationalInference: Abstract base class for variational inference. Specific

class WGANInference: Parameter estimation with GAN-style training (Goodfellow et al.,

Functions

complete_conditional(...): Returns the conditional distribution RandomVariable