pyhgf.utils.vectorised_belief_propagation.prediction_pass#

pyhgf.utils.vectorised_belief_propagation.prediction_pass(network, x)[source]#

Forward-only sweep through the network.

Sets the predictors on the top element and runs the top-down prediction sweep — no prediction errors, posterior updates, or weight learning — returning the bottom element’s expected_mean. Used by pyhgf.model.DeepNetwork.predict().

Parameters:

  • network (VectorisedNetwork) – The current vectorised network state.

  • x (Array) – The predictors set on the top element.

Returns:

The bottom element’s expected_mean after the forward sweep.

Return type:

expected_mean