descent_drift_field
def
The continuous-time descent drift vector field at memory weight H and scaled satisfaction ς with descent rate η and threshold G₀ is (-η * H, η * (ς - G₀)).
Statement
def descent_drift_field (η G₀ : ℝ) (p : ℝ × ℝ) : ℝ × ℝ- Defined in
- arXiv:2611.06076:808
- Verified
- 08 Nov 2026
- Axioms
- none listed
- Used by 1
- op_1172_optimal_constant_potential_function_existsThere exist an activation budget f ∈ (0, 1), descent parameters η > 0 and G₀ ≥ 1, a constant σ_star ∈ (1, 2) satisfying σ_star³ - 19 *…
Read back from the Lean
The function descent_drift_field, given real numbers η and G₀ and a point p = (p.1, p.2) : ℝ × ℝ, produces the pair of real numbers (-η * p.1, η * (p.2 - G₀)).
Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.
Lean source view module on GitHub
/-- Continuous-time descent drift vector field for memory weight H and satisfaction ς. -/ def descent_drift_field (η G₀ : ℝ) (p : ℝ × ℝ) : ℝ × ℝ := (-η * p.1, η * (p.2 - G₀))