Stability-controlled update fraction

Read the equation

The stability input is clipped to [0,1] and mapped to a gate between γ_min and 1. The next intermediate state takes the fraction γ of the gap from the current state to the proposed blended state. When γ_min lies in [0,1], this is a convex interpolation; a low gate slows response but does not prove the whole coupled system is stable.

Terms & definitions

How it works

Read the definition from its right-hand expression to the quantity on the left. Its surrounding source records the intended input, state, or parameter context. Where the source leaves an operation unspecified, this page preserves that gap.

Assumptions & limits

Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.

Original context

## 4.4 Stability Gating

```
γ = γ_min + (1 - γ_min) · clamp₀₁(stability)
Ψ'''_i(t+1) = Ψ_i(t) + γ · (Ψ''_i(t+1) - Ψ_i(t))
```

**Where**:
- `γ_min` = minimum gate value (default 0.1)
- `stability` = external stability signal

**Explanation**: When stability is low, updates are damped (γ approaches γ_min). This prevents erratic behavior during uncertain conditions.

## 4.5 Final Normalization

```
Ψ_i(t+1) = Ψ'''_i(t+1) · b(t+1)
```

**Where**:
- `b(t+1)` = normalization factor from equation 2.2

**Explanation**: All family activations are scaled by the normalizer, ensuring bounded behavior.

---

γ = γ_min + (1 - γ_min) · clamp₀₁(stability)
Ψ'''_i(t+1) = Ψ_i(t) + γ · (Ψ''_i(t+1) - Ψ_i(t))

Related builds