Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Maximal tail bound for the weighted noise sums in a fixed direction

Proved
StochApproxDyn.SubgaussianNoise.directional_maximal_tail

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

concentrationmartingalesp2o-batch-pfp1ap2o-gran-per-chapterp2o-plan-paperp2o-v1stochastic-approximationsub-gaussian

Let {xn}\{x_n\}{xn​} be a Robbins–Monro algorithm xn+1−xn=γn+1(F(xn)+Un+1)x_{n+1}-x_n=\gamma_{n+1}(F(x_n)+U_{n+1})xn+1​−xn​=γn+1​(F(xn​)+Un+1​) whose noise {Un}\{U_n\}{Un​} is subgaussian with constant Γ>0\Gamma>0Γ>0. Let τn=∑i=1nγi\tau_n=\sum_{i=1}^n\gamma_iτn​=∑i=1n​γi​ and m(t)=sup⁡{k≥0:τk≤t}m(t)=\sup\{k\ge0:\tau_k\le t\}m(t)=sup{k≥0:τk​≤t}. Fix a unit vector e∈Rde\in\mathbb R^de∈Rd, α>0\alpha>0α>0, n∈Nn\in\mathbb Nn∈N and T>0T>0T>0, and write N=m(τn+T)N=m(\tau_n+T)N=m(τn​+T). Then

P(sup⁡n<k≤N⟨e,∑i=nk−1γi+1Ui+1⟩≥α)≤exp⁡(−α22Γ∑i=nN−1γi+12).P\Big(\sup_{n<k\le N}\Big\langle e,\sum_{i=n}^{k-1}\gamma_{i+1}U_{i+1}\Big\rangle\ge\alpha\Big)\le\exp\Big(\frac{-\alpha^2}{2\Gamma\sum_{i=n}^{N-1}\gamma_{i+1}^2}\Big).P(n<k≤Nsup​⟨e,i=n∑k−1​γi+1​Ui+1​⟩≥α)≤exp(2Γ∑i=nN−1​γi+12​−α2​).

The paper uses this bound for the 2d2d2d directions e∈{±e1,…,±ed}e\in\{\pm e_1,\dots,\pm e_d\}e∈{±e1​,…,±ed​} of the canonical basis; combining them controls the norm of the noise sums over a time window of length TTT, which leads to estimate (18).

Formalization Note The event "the supremum over the finite range n<k≤Nn<k\le Nn<k≤N is at least α\alphaα" is written as "some kkk in that range has ⟨e,⋅⟩≥α\langle e,\cdot\rangle\ge\alpha⟨e,⋅⟩≥α"; an empty range gives the empty event. The statement is for every unit vector eee, which contains the paper's case. If ∑i=nN−1γi+12=0\sum_{i=n}^{N-1}\gamma_{i+1}^2=0∑i=nN−1​γi+12​=0 the intended right side is e−∞=0e^{-\infty}=0e−∞=0, but Lean's division by 000 gives e0=1e^{0}=1e0=1; the statement is then trivially true, and also true in the intended reading since every noise sum in the range vanishes.

Preamble
import Mathlib
import Definitions.Def_StochApproxDyn_MartingaleNoise_Interpolation
import Definitions.Def_StochApproxDyn_MartingaleNoise_AssumptionA1
import Definitions.Def_StochApproxDyn_MartingaleNoise_RobbinsMonro
import Definitions.Def_StochApproxDyn_SubgaussianNoise_Subgaussian

open MeasureTheory ProbabilityTheory Filter Topology
open scoped ENNReal InnerProductSpace
Formal statement
namespace StochApproxDyn.SubgaussianNoise

/-- Benaïm 1999, §4.2, proof of Proposition 4.4, p. 17: maximal tail bound in a fixed direction.
For a Robbins–Monro algorithm with subgaussian noise (constant `Γ`), every unit vector `e`,
every `α > 0`, every `n` and every `T > 0`,
`P(sup_{n < k ≤ m(τ_n+T)} ⟨e, ∑_{i=n}^{k-1} γ_{i+1} U_{i+1}⟩ ≥ α)
  ≤ exp(−α² / (2Γ ∑_{i=n}^{m(τ_n+T)-1} γ_{i+1}²))`.
The supremum over the finite range of `k` is at least `α` iff some term is; the event is
written that way (an empty range gives the empty event). When the block sum of `γ²` vanishes,
Lean's `x / 0 = 0` makes the right side `1`, so the bound is then trivially true (and the left
side is `0` anyway). The paper states it for `e ∈ {±e_1, …, ±e_d}`; every unit vector is
allowed here. -/
theorem directional_maximal_tail {d : ℕ} {Ω : Type*} {m0 : MeasurableSpace Ω}
    (P : Measure Ω) [IsProbabilityMeasure P] (ℱ : Filtration ℕ m0)
    (F : EuclideanSpace ℝ (Fin d) → EuclideanSpace ℝ (Fin d)) (γ : ℕ → ℝ)
    (x U : ℕ → Ω → EuclideanSpace ℝ (Fin d)) (hRM : StochApproxDyn.MartingaleNoise.IsRobbinsMonro P ℱ F γ x U)
    (Γ : ℝ) (hΓ : 0 < Γ) (hsg : IsSubgaussianWith P ℱ U Γ)
    (e : EuclideanSpace ℝ (Fin d)) (he : ‖e‖ = 1) (α : ℝ) (hα : 0 < α)
    (n : ℕ) (T : ℝ) (hT : 0 < T) :
    P {ω | ∃ k ∈ Finset.Ioc n (StochApproxDyn.MartingaleNoise.stepIndex γ (StochApproxDyn.Interpolation.tau γ n + T)),
        α ≤ ⟪e, ∑ i ∈ Finset.Ico n k, γ (i + 1) • U (i + 1) ω⟫_ℝ}
      ≤ ENNReal.ofReal (Real.exp (-α ^ 2 /
          (2 * Γ * ∑ i ∈ Finset.Ico n (StochApproxDyn.MartingaleNoise.stepIndex γ (StochApproxDyn.Interpolation.tau γ n + T)), γ (i + 1) ^ 2))) := by sorry

end StochApproxDyn.SubgaussianNoise
Source
Benaïm, Dynamics of Stochastic Approximation Algorithms, Séminaire de Probabilités XXXIII, LNM 1709 (1999), DOI 10.1007/BFb0096509, Section 4.2, proof of Proposition 4.4, p. 17 (PDF p. 18), unnumbered displays after 'Set R = α/(Γ Σ γ²_{i+1}), β = Rα and θ = Re'
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 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