Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 11.14: RepeatedHE is BIC if ε<13E[Gt⋅1{Gt>0}]\varepsilon < \frac13\mathbb{E}[G_t \cdot \mathbf 1\{G_t > 0\}]ε<31​E[Gt​⋅1{Gt​>0}] for each time t>N0t > N_0t>N0​

Proved
IntroBandits.repeatedHE_bic_of_gaps

by naimengye · Sep 24, 2026 · Mathlib 0df444a (Lean v4.33.1)

bayesian-incentive-compatibilityhidden-explorationincentivized-exploration

Corollary 11.14. RepeatedHE is BIC if ε<13 E[Gt⋅1{Gt>0}]\varepsilon < \frac13\,\mathbb{E}[G_t \cdot \mathbf 1\{G_t > 0\}]ε<31​E[Gt​⋅1{Gt​>0}] for each time t>N0t > N_0t>N0​, where Gt=E[μ2−μ1∣St]G_t = \mathbb{E}[\mu_2 - \mu_1 \mid S_t]Gt​=E[μ2​−μ1​∣St​] is the posterior gap given the exploration data StS_tSt​.

Formally: PPP a prior supported on the finite F⊆[0,1]2F \subseteq [0,1]^2F⊆[0,1]2 with μ10≥μ20\mu^0_1 \ge \mu^0_2μ10​≥μ20​, any reward family, any bandit algorithm ALG\mathrm{ALG}ALG, N0N_0N0​ initial rounds, ε>0\varepsilon > 0ε>0 and a horizon TTT. If for every round of the run after the initial ones (record lengths nnn with N0≤n<TN_0 \le n < TN0​≤n<T) ε<13E[max⁡(0,G)]\varepsilon < \frac13 \mathbb{E}[\max(0, G)]ε<31​E[max(0,G)], where GGG is the posterior gap given the exploration rounds of the first nnn rounds and the expectation is under the law of the first nnn rounds of RepeatedHE, then RepeatedHE is BIC (Definition 11.4) over the TTT rounds: for every round and every two distinct arms a,a′a, a'a,a′ with Pr⁡[rect=a]>0\Pr[\mathrm{rec}_t = a] > 0Pr[rect​=a]>0, E[(μa−μa′)1{rect=a}]≥0\mathbb{E}[(\mu_a - \mu_{a'})\mathbf 1\{\mathrm{rec}_t = a\}] \ge 0E[(μa​−μa′​)1{rect​=a}]≥0 under the law of the compliant run.

Preamble
import Definitions.Def_IntroBandits_Agents

open MeasureTheory ProbabilityTheory BanditAlgorithm
Formal statement
namespace IntroBandits

theorem repeatedHE_bic_of_gaps (P : Measure (Fin 2 → ℝ)) [IsProbabilityMeasure P]
    (F : Finset (Fin 2 → ℝ)) (hF : P (↑F)ᶜ = 0)
    (hunit : ∀ μ ∈ F, ∀ a, μ a ∈ Set.Icc (0 : ℝ) 1) (fam : RewardFamily)
    (hprior : priorMean P F 1 ≤ priorMean P F 0) (A : BanditPolicy 2) (N₀ : ℕ) {ε : ℝ}
    (hε : 0 < ε) {T : ℕ}
    (hgap : ∀ n, N₀ ≤ n → n < T →
      ε < heExpect P F fam A N₀ ε n (fun h ↦ max 0 (explGap P F fam h)) / 3) :
    IsBIC F (repeatedHELaw P F fam A N₀ ε T) recHE := by sorry

end IntroBandits
Source
Slivkins, Introduction to Multi-Armed Bandits, arXiv:1904.07272 (FnT ML 12, 2019), §11.4 p. 151, Corollary 11.14 (Lemma 11.10 applied round by round with signal S_t)
Human review
  • Endorsed by Shuze Chen · Sep 25, 2026

    Confirmed by the moderator at approval.

  • Endorsed by naimengye · Sep 25, 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