blackjax.smc.ess

blackjax.smc.ess#

All things related to SMC effective sample size

Functions#

ess(→ float | blackjax.types.Array)

Compute the effective sample size.

log_ess(→ float | blackjax.types.Array)

Compute the logarithm of the effective sample size.

ess_solver(→ float | blackjax.types.Array)

ESS solver for computing the next increment of SMC tempering.

Module Contents#

ess(log_weights: blackjax.types.Array) → float | blackjax.types.Array[source]#

Compute the effective sample size.

Parameters:

log_weights (Array) – Log-weights of the sample, shape (n_particles,).

Returns:

ess – The effective sample size.

Return type:

float | Array

log_ess(log_weights: blackjax.types.Array) → float | blackjax.types.Array[source]#

Compute the logarithm of the effective sample size.

Parameters:

log_weights (Array) – Log-weights of the sample, shape (n_particles,).

Returns:

log_ess – The logarithm of the effective sample size.

Return type:

float | Array

ess_solver(logdensity_fn: Callable, particles: blackjax.types.ArrayLikeTree, target_ess: float | blackjax.types.Array, max_delta: float | blackjax.types.Array, root_solver: Callable) → float | blackjax.types.Array[source]#

ESS solver for computing the next increment of SMC tempering.

Parameters:

  • logdensity_fn (Callable) – The log probability function we wish to sample from.

  • particles (ArrayLikeTree) – Current particles of the tempered SMC algorithm.

  • target_ess (float | Array) – Target effective sample size (ESS) for the next increment of SMC tempering.

  • max_delta (float | Array) – Maximum acceptable delta increment.

  • root_solver (Callable) – A solver to find the root of a function. Signature is root_solver(fun, min_delta, max_delta). Use e.g. dichotomy from blackjax.smc.solver.

Returns:

delta – The increment that solves for the target ESS.

Return type:

float | Array

Contents