Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 3, p. 5 — one-step expected gap contraction

Proved
KarimiPL.RCD.expected_cdStep_contraction

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

coordinate-descentexpectationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1polyak-lojasiewicz

Let d≥1d\ge1d≥1, L>0L>0L>0, and let fff obey the coordinate-wise upper bound (5). Suppose x∗x^*x∗ is a global minimizer, f∗=f(x∗)f^*=f(x^*)f∗=f(x∗), and fff satisfies the Polyak–Łojasiewicz (PL) inequality 12∥∇f(x)∥2≥μ(f(x)−f∗)\frac12\|\nabla f(x)\|^2\ge\mu(f(x)-f^*)21​∥∇f(x)∥2≥μ(f(x)−f∗) for some μ>0\mu>0μ>0. If III is uniform over the ddd coordinates, then for every xxx,

EI[f(TI(x))−f∗]≤(1−μdL)(f(x)−f∗).\mathbb E_I[f(T_I(x))-f^*] \le\left(1-\frac\mu{dL}\right)(f(x)-f^*).EI​[f(TI​(x))−f∗]≤(1−dLμ​)(f(x)−f∗).

This is the conditional one-step contraction used in Theorem 3. The minimizer fixes the optimal value; it is not an independently chosen lower bound.

Formalization Note The published PL predicate is written with both sides multiplied by two. The expectation is an explicit finite average.

Preamble
import Mathlib
import Definitions.Def_KarimiPL_RCD_Setting
import Definitions.Def_ShiOptrates_FunctionClass
Formal statement
namespace KarimiPL.RCD

/-- Conditional expected contraction in the proof of Theorem 3, p. 5. -/
theorem expected_cdStep_contraction {d : ℕ} (hd : 0 < d)
    (f : KarimiPL.Rel.E d → ℝ) (L : ℝ) (hL : 0 < L) (hcoord : CoordLipschitz f L)
    (xstar : KarimiPL.Rel.E d) (hxstar : ∀ y, f xstar ≤ f y)
    (mu : ℝ) (hmu : 0 < mu)
    (hPL : ShiOptRates.PolyakLojasiewicz f (gradient f) mu (f xstar)) :
    ∀ x : KarimiPL.Rel.E d,
      (1 / (d : ℝ)) * ∑ i : Fin d, (f (cdStep f L x i) - f xstar) ≤
        (1 - mu / ((d : ℝ) * L)) * (f x - f xstar) := by sorry

end KarimiPL.RCD
Source
Karimi, Nutini, Schmidt, arXiv:1608.04636v4, proof of Theorem 3, final displayed inequality, p. 5
Human review
  • Endorsed by Shuze Chen · Oct 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 9, 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