Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Exp4 expected potential bound

Proved
BanditAlgorithm.exp4_expected_potential_bound

by Harry_Xu · Jul 29, 2026 · Mathlib c5ea003 (Lean v4.30.0)

bandit-algorithmsonline-learning

Let k≥1k\ge1k≥1, M≥2M\ge2M≥2, n≥1n\ge1n≥1, and η>0\eta>0η>0. Let rewards satisfy xt,a∈[0,1]x_{t,a}\in[0,1]xt,a​∈[0,1], let every expert advice row (Et,m,a)a(E_{t,m,a})_a(Et,m,a​)a​ be a probability distribution, and suppose the learner follows Exp4 with exploration parameter γ=0\gamma=0γ=0. Write Qt,mQ_{t,m}Qt,m​ for the exponential weight of expert mmm, X~t,m\widetilde X_{t,m}Xt,m​ for its one-round importance-weighted score, and

Vn=∑t=1n∑j=1MQt,j(1−X~t,j)2.V_n=\sum_{t=1}^{n}\sum_{j=1}^{M}Q_{t,j}(1-\widetilde X_{t,j})^2.Vn​=t=1∑n​j=1∑M​Qt,j​(1−Xt,j​)2.

For every fixed expert mmm,

E[S~n,m]−E ⁣[∑t=1n∑j=1MQt,jX~t,j]≤log⁡Mη+η2 E[Vn].\mathbb E[\widetilde S_{n,m}] -\mathbb E\!\left[\sum_{t=1}^{n}\sum_{j=1}^{M}Q_{t,j}\widetilde X_{t,j}\right] \le \frac{\log M}{\eta}+\frac{\eta}{2}\,\mathbb E[V_n].E[Sn,m​]−E[t=1∑n​j=1∑M​Qt,j​Xt,j​]≤ηlogM​+2η​E[Vn​].

This is the expected form of the exponential-weights potential inequality used in Exp4, before the mixture score and quadratic term are evaluated.

Formalization Note exp4MixtureEstimate is the cumulative double sum and exp4QuadraticMass is VnV_nVn​; all expectations are integrals over the adversarial interaction measure.

Preamble
import Definitions.Def_ContextualAdversarialBandit
import Definitions.Def_exp4Analysis

open MeasureTheory ProbabilityTheory
Formal statement
theorem BanditAlgorithm.exp4_expected_potential_bound
    {k M : ℕ} (hk : 0 < k) (hM : 1 < M) (n : ℕ) (hn : 0 < n)
    (η : ℝ) (hη : 0 < η)
    (x : ℕ → Fin k → ℝ) (hx : ∀ t : ℕ, ∀ a : Fin k, x t a ∈ Set.Icc (0 : ℝ) 1)
    (E : ℕ → Fin M → Fin k → ℝ)
    (hE0 : ∀ t : ℕ, ∀ m : Fin M, ∀ a : Fin k, 0 ≤ E t m a)
    (hE1 : ∀ t : ℕ, ∀ m : Fin M, ∑ a, E t m a = 1)
    (π : BanditPolicy k)
    (hπ : IsExp4Policy η 0 E π)
    (m : Fin M) :
    (∫ h, exp4Estimate η 0 E n h m ∂(adversarialMeasure x π n)) -
        (∫ h, exp4MixtureEstimate η E n h ∂(adversarialMeasure x π n)) ≤
      Real.log M / η +
        (η / 2) *
          (∫ h, exp4QuadraticMass η E n h ∂(adversarialMeasure x π n)) := by
  sorry
Source
Lattimore and Szepesvári, Bandit Algorithms (Cambridge University Press, 2020), Lemma 18.2 and Theorem 18.1 proof, printed p. 230, especially Eq. (18.9). https://tor-lattimore.com/downloads/book/book.pdf

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