Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eq. (2.4) — pseudo-regret of α\alphaα-UCB for [0,1][0,1][0,1] rewards

Proved
RegretBandits.Stochastic.alpha_ucb_pseudo_regret

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

banditsp2o-batch-pfp1bp2o-gran-per-chapterp2o-plan-bookp2o-v1regretucb

Consider a stochastic bandit with K≥2K\ge2K≥2 arms whose rewards take values in [0,1][0,1][0,1], with means μi\mu_iμi​, μ∗=max⁡iμi\mu^*=\max_i\mu_iμ∗=maxi​μi​ and gaps Δi=μ∗−μi\Delta_i=\mu^*-\mu_iΔi​=μ∗−μi​. Taking ψ(λ)=λ2/8\psi(\lambda)=\lambda^2/8ψ(λ)=λ2/8 in (2.2) (Hoeffding's lemma) gives ψ∗(ε)=2ε2\psi^*(\varepsilon)=2\varepsilon^2ψ∗(ε)=2ε2, and (α,ψ)(\alpha,\psi)(α,ψ)-UCB with this ψ\psiψ is called α\alphaα-UCB. For α>2\alpha>2α>2 and every horizon nnn,

R‾n≤∑i:Δi>0(2αΔiln⁡n+αα−2).\overline R_n\le\sum_{i:\Delta_i>0}\left(\frac{2\alpha}{\Delta_i}\ln n+\frac{\alpha}{\alpha-2}\right).Rn​≤i:Δi​>0∑​(Δi​2α​lnn+α−2α​).

This is the most used special case of Theorem 2.1, for bounded rewards; the matching lower bound in its dependence on ln⁡n/Δi\ln n/\Delta_ilnn/Δi​ is Theorem 2.2 combined with (2.8).

Formalization Note. The algorithm is run with ψ(λ)=λ2/8\psi(\lambda)=\lambda^2/8ψ(λ)=λ2/8 and plays every arm once before using the index. Rewards lie in [0,1][0,1][0,1] almost surely. Here Δi≤1\Delta_i\le1Δi​≤1, so the printed constant α/(α−2)\alpha/(\alpha-2)α/(α−2) is correct as stated.

Preamble
import Mathlib
import Definitions.Def_ImprovedLinBandits_UCBDelta_armModel
import Definitions.Def_RegretBandits_Stochastic_model
import Definitions.Def_RegretBandits_Stochastic_alphaPsiUCB
Formal statement
namespace RegretBandits.Stochastic

open MeasureTheory ImprovedLinBandits.UCBDelta

/-- Eq. (2.4) of Bubeck and Cesa-Bianchi (arXiv:1204.5721v2, p. 11): for `[0, 1]`-valued rewards,
α-UCB, i.e. (α, ψ)-UCB with `ψ(λ) = λ²/8` (Hoeffding's lemma, `ψ*(ε) = 2ε²`) and `α > 2`, satisfies
`R̄_n ≤ ∑_{i : Δ_i > 0} ((2α/Δ_i) ln n + α/(α - 2))`. -/
theorem alpha_ucb_pseudo_regret {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω)
    [IsProbabilityMeasure P] {K : ℕ} (hK : 2 ≤ K) (X : Fin K → ℕ → Ω → ℝ) (μ : Fin K → ℝ)
    (hX : IsStochasticBandit P X μ)
    (hbdd : ∀ i k, ∀ᵐ ω ∂P, X i k ω ∈ Set.Icc (0 : ℝ) 1)
    (α : ℝ) (hα : 2 < α) (I : ℕ → Ω → Fin K) (hI : ∀ t, Measurable (I t))
    (hrun : IsAlphaPsiUCBRun (fun l => l ^ 2 / 8) α X I) (n : ℕ) :
    pseudoRegretBar P μ I n ≤
      ∑ i ∈ Finset.univ.filter (fun i => 0 < gap μ i),
        (2 * α / gap μ i * Real.log (n : ℝ) + α / (α - 2)) := by sorry

end RegretBandits.Stochastic
Source
Bubeck, Cesa-Bianchi, Regret Analysis of Stochastic and Nonstochastic Multi-armed Bandit Problems, arXiv:1204.5721v2, p. 11, Eq. (2.4)
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