pyhgf.updates.vectorised.continuous.vectorised_continuous_posterior_update#

pyhgf.updates.vectorised.continuous.vectorised_continuous_posterior_update(layer, value_child=None, volatility_child=None, volatility_updates='eHGF', time_step=1.0, max_posterior_precision=10000000000.0, mean_field_updates=False)[source]#

Dispatch the continuous posterior update on the layer’s children.

Mirrors the assignment rule of pyhgf.utils.get_update_sequence.get_update_sequence(): layers without a volatility child always take the standard update; layers with one take the update named by volatility_updates.

Parameters:

  • layer (LayerState) – The parent layer’s state (being updated).

  • value_child (ValueChild | None) – The value child, or None.

  • volatility_child (VolatilityChild | None) – The volatility child, or None.

  • volatility_updates (str) – One of "standard", "eHGF", or "unbounded".

  • time_step (float) – Current time step.

  • max_posterior_precision (float) – Upper bound applied to the posterior precision write.

  • mean_field_updates (bool) – If True, the value-coupling messages use the original mean-field forms and the variance reconstructions drop the MGF term the mean-field prediction never added; see the individual update schemes.

Returns:

Updated layer state with posterior precision and mean.

Return type:

LayerState