Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3, p. 16 — NAG-SC with s = 1/(4L): f(xₖ) − f(x⋆) ≤ 5L‖x₀ − x⋆‖²/(1 + √(μ/L)/12)ᵏ

Proved
HighResODE.NAGSC.theorem_3

by mikedeng1 · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

accelerated-gradientlinear-convergencenesterovp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1strongly-convex

Theorem 3 (Convergence of NAG-SC). Let f∈Sμ,L1(Rn)f\in\mathcal S^1_{\mu,L}(\mathbb R^n)f∈Sμ,L1​(Rn), that is, fff is differentiable and convex with an LLL-Lipschitz gradient (L>0L>0L>0) and is μ\muμ-strongly convex with 0<μ≤L0<\mu\le L0<μ≤L, and let x⋆x^\starx⋆ be its minimizer. Run Nesterov's accelerated gradient method NAG-SC (1.3) with step size s=1/(4L)s=1/(4L)s=1/(4L) from an arbitrary x0=y0x_0=y_0x0​=y0​:

yk+1=xk−s∇f(xk),xk+1=yk+1+1−μs1+μs(yk+1−yk).y_{k+1}=x_k-s\nabla f(x_k),\qquad x_{k+1}=y_{k+1}+\frac{1-\sqrt{\mu s}}{1+\sqrt{\mu s}}(y_{k+1}-y_k).yk+1​=xk​−s∇f(xk​),xk+1​=yk+1​+1+μs​1−μs​​(yk+1​−yk​).

Then, for all k≥0k\ge0k≥0,

f(xk)−f(x⋆)≤5L∥x0−x⋆∥2(1+112μ/L)k.f(x_k)-f(x^\star)\le\frac{5L\|x_0-x^\star\|^2}{\big(1+\frac1{12}\sqrt{\mu/L}\big)^k}.f(xk​)−f(x⋆)≤(1+121​μ/L​)k5L∥x0​−x⋆∥2​.

The bound says log⁡(f(xk)−f(x⋆))≤−Ω(kμ/L)\log(f(x_k)-f(x^\star))\le-\Omega(k\sqrt{\mu/L})log(f(xk​)−f(x⋆))≤−Ω(kμ/L​), the optimal accelerated rate for first-order methods on smooth strongly convex functions; the paper obtains it from a single discrete Lyapunov function derived from the high-resolution ODE of NAG-SC.

Formalization Note Rn\mathbb R^nRn is EuclideanSpace ℝ (Fin n) and ∇f\nabla f∇f is Mathlib's gradient f. The minimizer x⋆x^\starx⋆ is a point xs with ∀ z, f xs ≤ f z. The NAG-SC run is the predicate IsNAGSC f μ (1 / (4 * L)) x y, which fixes both sequences from x0x_0x0​; iterates are indexed from 000, and the bound is claimed for every kkk, including k=0k=0k=0. The step size is exactly 1/(4L)1/(4L)1/(4L), as in the theorem.

Preamble
import Mathlib
import Definitions.Def_HighResODE_NAGSC_FunctionClass
import Definitions.Def_HighResODE_NAGSC_Setting
Formal statement
namespace HighResODE.NAGSC

open scoped InnerProductSpace

/-- Theorem 3 (Convergence of NAG-SC), p. 16: for `f ∈ S¹_{μ,L}(ℝⁿ)` and step size
`s = 1/(4L)`, the NAG-SC iterates satisfy
`f(x_k) − f(x⋆) ≤ 5L‖x₀ − x⋆‖² / (1 + (1/12)√(μ/L))^k` for all `k ≥ 0`. -/
theorem theorem_3 {n : ℕ} (f : E n → ℝ) (μ L : ℝ) (hf : IsS1 f μ L) (xs : E n)
    (hmin : ∀ z : E n, f xs ≤ f z) (x y : ℕ → E n)
    (hrun : IsNAGSC f μ (1 / (4 * L)) x y) :
    ∀ k : ℕ, f (x k) - f xs ≤
      5 * L * ‖x 0 - xs‖ ^ 2 / (1 + 1 / 12 * Real.sqrt (μ / L)) ^ k := by sorry

end HighResODE.NAGSC
Source
Shi, Du, Jordan & Su, Understanding the Acceleration Phenomenon via High-Resolution Differential Equations, arXiv:1810.08907v3, p. 16, Theorem 3
Human review
  • Endorsed by Shuze Chen · Oct 10, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 10, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me