AIM Framework glossary

Network Loop Gain

Canonical definition

Network Loop Gain (GnetG{net}). The Network Loop Gain is the cross-Individual contribution to the total stability condition of the three-layer formalism, decomposing additively as Gnet=GnetΛ+GnetKG{net} = G{net}^{\Lambda} + G{net}^{K}, where GnetΛ=α⋅β⋅ρ(Λ)⋅σ′(V∗)G{net}^{\Lambda} = \alpha\cdot \beta\cdot \rho(\Lambda)\cdot \sigma'(V^{*}) is the fast-channel network gain (with ρ(Λ)\rho(\Lambda) the spectral radius of the Observability matrix) and GnetK=ρ(Koff)⋅∥∂ℓ/∂E∥⋅β ∥∂μ/∂b∥⋅ω(Tint)‾G{net}^{K} = \rho(K{off})\cdot \|\partial\ell/\partial E\|\cdot \beta\,\|\partial\mu/\partial b\|\cdot \overline{\omega(T{int})} is the slow-channel network gain (with ρ(Koff)\rho(K{off}) the spectral radius of the off-diagonal Evidence-coupling matrix and ω(Tint)‾\overline{\omega(T{int})} the population-averaged Evidence-emission coefficient).

Explanation

The Network Loop Gain is the population-scale analogue of the within-Individual Internal Loop Gain and the Slow-Layer Loop Gain. It captures how strongly an Individual's valuation and Belief outputs feed back into the system through other Individuals, via two architecturally distinct channels: the Observability channel (one Individual's behaviour becomes another's Model signal) and the Evidence channel (one Individual's emitted Evidence proposition enters another's Belief Register). The additive decomposition makes each channel separately diagnosable and separately interveneable, which is the architectural basis for the five-channel Knowledge Patch footprint at the network scale.

Epistemic status

Framework-Specific Construct. Source: Appendix A.5.7.4.

Derivation

Appendix A, Section A.5.7.4; Chapter 16, Section 16.19.

Related terms