Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3 — the SA inner loop (SA_{i,k}) reaches E[∥zi,t−x^i(yk)∥2∣Fk]≤Qi/(t+1)\mathbb E[\|z_{i,t}-\hat x_i(y_k)\|^2\mid\mathcal F_k]\le Q_i/(t+1)E[∥zi,t​−x^i​(yk​)∥2∣Fk​]≤Qi​/(t+1)

Proved
LeiBR.Sync.lemma_3

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

best-responseconvergence-ratep2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-approximation

Assume Assumption 1 and μ>0\mu>0μ>0, and let ΠXi\Pi_{X_i}ΠXi​​ be the Euclidean projection onto XiX_iXi​. Run Algorithm 1 with the inner loop (SAi,k_{i,k}i,k​) from a deterministic x0∈Xx_0\in Xx0​∈X, with any step counts ji,kj_{i,k}ji,k​ and measurable samples ξi,kt\xi^t_{i,k}ξi,kt​; the sampled gradient map (x,s)↦∇xiψi(x;s)(x,s)\mapsto\nabla_{x_i}\psi_i(x;s)(x,s)↦∇xi​​ψi​(x;s) is measurable. Fix a major iteration kkk and a player iii, and suppose that for every t=1,…,ji,kt=1,\dots,j_{i,k}t=1,…,ji,k​ the sampled gradient Gt=∇xiψi(zi,t,y−i,k;ξi,kt)G_t=\nabla_{x_i}\psi_i(z_{i,t},y_{-i,k};\xi^t_{i,k})Gt​=∇xi​​ψi​(zi,t​,y−i,k​;ξi,kt​) satisfies E∥Gt∥2<∞\mathbb E\|G_t\|^2<\inftyE∥Gt​∥2<∞,

E[Gt ∣ σ{Fk,ξi,k[t−1]}]=∇xifi(zi,t,y−i,k),E[∥Gt∥2 ∣ σ{Fk,ξi,k[t−1]}]≤Mi2a.s.\mathbb E\big[G_t\,\big|\,\sigma\{\mathcal F_k,\xi^{[t-1]}_{i,k}\}\big]=\nabla_{x_i}f_i(z_{i,t},y_{-i,k}),\qquad \mathbb E\big[\|G_t\|^2\,\big|\,\sigma\{\mathcal F_k,\xi^{[t-1]}_{i,k}\}\big]\le M_i^2\quad\text{a.s.}E[Gt​​σ{Fk​,ξi,k[t−1]​}]=∇xi​​fi​(zi,t​,y−i,k​),E[∥Gt​∥2​σ{Fk​,ξi,k[t−1]​}]≤Mi2​a.s.

Let Qi=2Mi2/μ2+2DXi2Q_i=2M_i^2/\mu^2+2D_{X_i}^2Qi​=2Mi2​/μ2+2DXi​2​. Then for every t=1,…,ji,kt=1,\dots,j_{i,k}t=1,…,ji,k​,

E[∥zi,t−x^i(yk)∥2 ∣ Fk]≤Qit+1a.s.\mathbb E\big[\|z_{i,t}-\hat x_i(y_k)\|^2\,\big|\,\mathcal F_k\big]\le\frac{Q_i}{t+1}\quad\text{a.s.}E[∥zi,t​−x^i​(yk​)∥2​Fk​]≤t+1Qi​​a.s.

This is the standard O(1/t)O(1/t)O(1/t) rate of stochastic approximation for a μ\muμ-strongly convex problem, applied to each player's proximal BR problem; it tells how many SA steps achieve the accuracy (8).

Formalization Note Two repairs of the printed hypotheses. (i) The paper writes ψi\psi_iψi​ where it means the sampled gradient ∇xiψi\nabla_{x_i}\psi_i∇xi​​ψi​. (ii) Its second hypothesis, read literally, says E[∥Gt∥2∣⋅]=∥∇xifi∥2\mathbb E[\|G_t\|^2\mid\cdot]=\|\nabla_{x_i}f_i\|^2E[∥Gt​∥2∣⋅]=∥∇xi​​fi​∥2, which would force zero sampling noise; the proof (p. 11) uses only the bound E[∥Gt∥2∣⋅]≤Mi2\mathbb E[\|G_t\|^2\mid\cdot]\le M_i^2E[∥Gt​∥2∣⋅]≤Mi2​, which is what is assumed here. It follows from the literal hypothesis, Assumption 1(d) and Jensen, so the statement is no weaker than the printed one. Fk\mathcal F_kFk​ and σ{Fk,ξi,k[t−1]}\sigma\{\mathcal F_k,\xi^{[t-1]}_{i,k}\}σ{Fk​,ξi,k[t−1]​} are generated by the samples; yk=xky_k=x_kyk​=xk​.

Preamble
import Mathlib
import Definitions.Def_SpectralProjGrad_Shared_IsProjOnto
import Definitions.Def_LeiBR_Sync_NashGame
import Definitions.Def_LeiBR_Sync_StochGame
import Definitions.Def_LeiBR_Sync_Algorithm1

open MeasureTheory
Formal statement
namespace LeiBR.Sync

/-- Lemma 3, p. 10 (hypotheses repaired, see the statement): in Algorithm 1 with the inner loop
(SA_{i,k}), fix a major iteration `k` and a player `i`, and suppose that for every
`t = 1, …, j_{i,k}` the sampled gradient `G = ∇_{x_i}ψ_i(z_{i,t}, y_{-i,k}; ξ^t_{i,k})` has
`E‖G‖² < ∞`, `E[G | σ{F_k, ξ^{[t-1]}_{i,k}}] = ∇_{x_i} f_i(z_{i,t}, y_{-i,k})` a.s. and
`E[‖G‖² | σ{F_k, ξ^{[t-1]}_{i,k}}] ≤ M_i²` a.s. Then
`E[‖z_{i,t} - x̂_i(y_k)‖² | F_k] ≤ Q_i/(t+1)` a.s. for every `t = 1, …, j_{i,k}`, where
`Q_i = 2M_i²/μ² + 2D²_{X_i}`. -/
theorem lemma_3
    {N : ℕ} {n : Fin N → ℕ} {Ω : Type*} [mΩ : MeasurableSpace Ω] (P : Measure Ω)
    [IsProbabilityMeasure P] {d : ℕ} (ξ : Ω → EuclideanSpace ℝ (Fin d))
    (X : ∀ i : Fin N, Set (Strat n i)) (ψ : Fin N → Profile n → EuclideanSpace ℝ (Fin d) → ℝ)
    (gψ : ∀ i : Fin N, Profile n → EuclideanSpace ℝ (Fin d) → Strat n i) (M : Fin N → ℝ)
    (hA1 : Assumption1 P ξ X ψ gψ M) (μ : ℝ) (hμ : 0 < μ)
    (xhat : Profile n → Profile n) (hxhat : IsProxBR X (payoff P ξ ψ) μ xhat)
    (proj : ∀ i : Fin N, Strat n i → Strat n i)
    (hproj : ∀ i : Fin N, SpectralProjGrad.Shared.IsProjOnto (X i) (proj i))
    (hgψ : ∀ i : Fin N,
      Measurable (fun p : Profile n × EuclideanSpace ℝ (Fin d) => gψ i p.1 p.2))
    (j : Fin N → ℕ → ℕ) (ξs : ℕ → Fin N → ℕ → Ω → EuclideanSpace ℝ (Fin d))
    (hξs : ∀ k (i : Fin N) t, Measurable (ξs k i t))
    (x0 : Profile n) (hx0 : x0 ∈ stratSet X)
    (k : ℕ) (i : Fin N)
    (hint : ∀ t : ℕ, 1 ≤ t → t ≤ j i k →
      Integrable (fun ω => ‖gψ i (Function.update (syncSA proj gψ μ j ξs x0 k ω) i
        (saZ proj gψ μ j ξs x0 k i t ω)) (ξs k i t ω)‖ ^ 2) P)
    (hunb : ∀ t : ℕ, 1 ≤ t → t ≤ j i k →
      P[fun ω => gψ i (Function.update (syncSA proj gψ μ j ξs x0 k ω) i
          (saZ proj gψ μ j ξs x0 k i t ω)) (ξs k i t ω) | innerField j ξs k i t]
        =ᵐ[P] fun ω => partialGrad (payoff P ξ ψ) i
          (Function.update (syncSA proj gψ μ j ξs x0 k ω) i (saZ proj gψ μ j ξs x0 k i t ω)))
    (hvar : ∀ t : ℕ, 1 ≤ t → t ≤ j i k →
      P[fun ω => ‖gψ i (Function.update (syncSA proj gψ μ j ξs x0 k ω) i
          (saZ proj gψ μ j ξs x0 k i t ω)) (ξs k i t ω)‖ ^ 2 | innerField j ξs k i t]
        ≤ᵐ[P] fun _ => M i ^ 2) :
    ∀ t : ℕ, 1 ≤ t → t ≤ j i k →
      P[fun ω => ‖saZ proj gψ μ j ξs x0 k i t ω - xhat (syncSA proj gψ μ j ξs x0 k ω) i‖ ^ 2 |
          infoField j ξs k] ≤ᵐ[P] fun _ => Qconst X M μ i / ((t : ℝ) + 1) := by sorry

end LeiBR.Sync
Source
Lei, Shanbhag, Pang & Sen, On Synchronous, Asynchronous, and Randomized Best-Response Schemes for Stochastic Nash Games, arXiv:1704.04578v2, p. 10, Lemma 3 (proof pp. 10–11)
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 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