morphological_silo (faber_neuroevolution v1.2.4)
View SourceMorphological Silo - Network structure and complexity management.
Part of the Liquid Conglomerate v2 architecture. The Morphological Silo manages: Network size constraints (neurons, connections) Complexity tracking and penalties Pruning thresholds Parameter efficiency optimization Sensor/actuator addition rates
Time Constant
τ = 30 (medium adaptation for structural changes)
Cross-Silo Signals
Outgoing: complexity_signal to task: Current network complexity level size_budget to resource: Network size requirements efficiency_score to economic: Parameter efficiency metric growth_stage to developmental: Structural development stage
Incoming: pressure_signal from resource: Resource constraint complexity_target from task: Target complexity level efficiency_requirement from economic: Efficiency targets expression_cost from regulatory: Cost of gene expression
Summary
Functions
-spec record_network_size(pid(), term(), non_neg_integer(), non_neg_integer()) -> ok.
-spec reset(pid()) -> ok.