Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 15.12, pp. 248–249 — player regrets bound the averaged duality gap

Proved
ModernOnlineLearning.SaddlePoint.theorem_15_12

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

duality-gaponline-learningp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1saddle-point

Let X,YX,YX,Y be nonempty closed convex sets, and let xt∈Xx_t\in Xxt​∈X, yt∈Yy_t\in Yyt​∈Y be the simultaneous outputs of Algorithm 15.1 for t=1,…,Tt=1,\ldots,Tt=1,…,T, where T≥1T\ge1T≥1. Player XXX receives ℓt(x)=f(x,yt)\ell_t(x)=f(x,y_t)ℓt​(x)=f(x,yt​) and player YYY receives ht(y)=−f(xt,y)h_t(y)=-f(x_t,y)ht​(y)=−f(xt​,y). Their regret identities say, for u∈Xu\in Xu∈X and v∈Yv\in Yv∈Y,

1T∑tf(xt,yt)−1T∑tf(u,yt)=1T∑t[f(xt,yt)−f(u,yt)],\frac1T\sum_t f(x_t,y_t)-\frac1T\sum_t f(u,y_t)=\frac1T\sum_t[f(x_t,y_t)-f(u,y_t)],T1​t∑​f(xt​,yt​)−T1​t∑​f(u,yt​)=T1​t∑​[f(xt​,yt​)−f(u,yt​)], 1T∑tf(xt,v)−1T∑tf(xt,yt)=1T∑t[f(xt,v)−f(xt,yt)].\frac1T\sum_t f(x_t,v)-\frac1T\sum_t f(x_t,y_t)=\frac1T\sum_t[f(x_t,v)-f(x_t,y_t)].T1​t∑​f(xt​,v)−T1​t∑​f(xt​,yt​)=T1​t∑​[f(xt​,v)−f(xt​,yt​)].

If fff is convex in its first argument and concave in its second, and uT′u_T'uT′​ minimizes f(⋅,yˉT)f(\cdot,\bar y_T)f(⋅,yˉ​T​) while vT′v_T'vT′​ maximizes f(xˉT,⋅)f(\bar x_T,\cdot)f(xˉT​,⋅) on their respective sets, then

f(xˉT,vT′)−f(uT′,yˉT)≤Regret⁡TX(uT′)+Regret⁡TY(vT′)T.f(\bar x_T,v_T')-f(u_T',\bar y_T)\le\frac{\operatorname{Regret}^X_T(u_T')+\operatorname{Regret}^Y_T(v_T')}{T}.f(xˉT​,vT′​)−f(uT′​,yˉ​T​)≤TRegretTX​(uT′​)+RegretTY​(vT′​)​.

The inequality converts two online regret bounds into a bound on the attained duality gap.

Formalization Note The two regret expressions are expanded as finite sums. The attained extrema appear as point witnesses, and the average uses the book's one-based index.

Preamble
import Mathlib
import Definitions.Def_ModernOnlineLearning_SaddlePoint_Setting

set_option autoImplicit false
Formal statement
namespace ModernOnlineLearning.SaddlePoint

/-- Theorem 15.12, pp. 248–249: the two regret identities and the
online-to-saddle-point duality-gap inequality. -/
theorem theorem_15_12 {EX EY : Type*}
    [NormedAddCommGroup EX] [NormedSpace ℝ EX] [FiniteDimensional ℝ EX]
    [NormedAddCommGroup EY] [NormedSpace ℝ EY] [FiniteDimensional ℝ EY]
    (X : Set EX) (Y : Set EY) (f : EX → EY → ℝ)
    (x : ℕ → EX) (y : ℕ → EY) (T : ℕ)
    (hT : 1 ≤ T) (hXne : X.Nonempty) (hYne : Y.Nonempty)
    (hXclosed : IsClosed X) (hYclosed : IsClosed Y)
    (hX : Convex ℝ X) (hY : Convex ℝ Y)
    (hx : ∀ t ∈ Finset.Icc 1 T, x t ∈ X)
    (hy : ∀ t ∈ Finset.Icc 1 T, y t ∈ Y)
    (hfx : ∀ v ∈ Y, ConvexOn ℝ X (fun w => f w v))
    (hfy : ∀ u ∈ X, ConcaveOn ℝ Y (f u))
    (xm : EX) (ym : EY)
    (hmin : IsMinOn X (fun w => f w (avg T y)) xm)
    (hmax : IsMaxOn Y (f (avg T x)) ym) :
    (∀ u ∈ X,
      (∑ t ∈ Finset.Icc 1 T, f (x t) (y t)) / T -
        (∑ t ∈ Finset.Icc 1 T, f u (y t)) / T =
        (∑ t ∈ Finset.Icc 1 T, (f (x t) (y t) - f u (y t))) / T) ∧
    (∀ v ∈ Y,
      (∑ t ∈ Finset.Icc 1 T, f (x t) v) / T -
        (∑ t ∈ Finset.Icc 1 T, f (x t) (y t)) / T =
        (∑ t ∈ Finset.Icc 1 T, (f (x t) v - f (x t) (y t))) / T) ∧
    (f (avg T x) ym - f xm (avg T y) ≤
      ((∑ t ∈ Finset.Icc 1 T, (f (x t) (y t) - f xm (y t))) +
        (∑ t ∈ Finset.Icc 1 T, (f (x t) ym - f (x t) (y t)))) / T) := by sorry

end ModernOnlineLearning.SaddlePoint
Source
Orabona, arXiv:1912.13213v10, Theorem 15.12, pp. 248–249
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).

  • Endorsed by iris · Nov 6, 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