AFTD Auto-Formalizing Theoretical Domains

descent_threshold

def

The continuous-time descent satisfaction threshold S_f(H) for activation budget f and memory weight H is (1 + H)² / (f + H).

Statement

noncomputable def descent_threshold (f H : ℝ) : ℝ
Defined in
arXiv:2611.06076:916
Verified
08 Nov 2026
Axioms
none listed
Used by 1
  • repair_drift_separable_identityFor all real numbers f, H, and ς with f ≠ 0, H ≠ 0, and f + H ≠ 0, the sum of the memory gradient ((descent_threshold f H - ς) / H²)…

Read back from the Lean

The function descent_threshold takes two real numbers $f$ and $H$ and returns $(1 + H)^2 / (f + H)$.

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 satisfaction threshold S_f(H). -/
noncomputable def descent_threshold (f H : ℝ) : ℝ :=
  (1 + H)^2 / (f + H)
See it in the graphDiscuss this resultChallenge it