Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section 3.4, Eq. (5) — size of the lifted class: log |A_{M,ρ}| ≤ C M log(e/ρ)

Proved
CalibratedMA.Randomized.eq_5

by mikedeng1 · Nov 10, 2026 · Mathlib 0df444a (Lean v4.33.1)

calibrationcountingp2o-batch-stoc27ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let ρ>0\rho>0ρ>0 with 1/ρ1/\rho1/ρ an integer, Kρ=∣Λρ∣=1/ρ+1K_\rho=|\Lambda_\rho|=1/\rho+1Kρ​=∣Λρ​∣=1/ρ+1, and let M≥1M\ge1M≥1 be an integer. The lifted test class AM,ρ\mathcal A_{M,\rho}AM,ρ​ of Section 3.4 satisfies the counting bound

∣AM,ρ∣≤∑L=1M(KρL+1)2L+1≤(M+1)(2Kρ)M+1,|\mathcal A_{M,\rho}|\le\sum_{L=1}^M\binom{K_\rho}{L+1}2^{L+1}\le (M+1)(2K_\rho)^{M+1},∣AM,ρ​∣≤L=1∑M​(L+1Kρ​​)2L+1≤(M+1)(2Kρ​)M+1,

and, for a universal constant CCC,

log⁡∣AM,ρ∣≤C Mlog⁡(e/ρ).(5)\log|\mathcal A_{M,\rho}|\le C\,M\log(e/\rho).\tag{5}log∣AM,ρ​∣≤CMlog(e/ρ).(5)

This bound makes the online learner's guarantee (Lemma 3.1) over ΦG,A\Phi_{\mathcal G,\mathcal A}ΦG,A​ cost only O(mlog⁡(e/ρ))O(m\log(e/\rho))O(mlog(e/ρ)) in the logarithm, which yields the (log⁡(2n/δ)/n)2/5(\log(2n/\delta)/n)^{2/5}(log(2n/δ)/n)2/5 term of Theorem 4.1.

Formalization Note. ρ=1/K\rho=1/Kρ=1/K with K≥1K\ge1K≥1 a natural number, so Kρ=K+1K_\rho=K+1Kρ​=K+1 and e/ρe/\rhoe/ρ is written exp 1 / (1/K). ∣AM,ρ∣|\mathcal A_{M,\rho}|∣AM,ρ​∣ is the cardinality of the Finset of distinct lifted functions. CCC is quantified before KKK and MMM.

Differences from the source:

  • encoding: ρ>0\rho>0ρ>0 with 1/ρ1/\rho1/ρ an integer → K∈NK\in\mathbb NK∈N, K≥1K\ge1K≥1, ρ=1/K\rho=1/Kρ=1/K; reason: same instances.
  • encoding: AM,ρ\mathcal A_{M,\rho}AM,ρ​ as a class of functions on [0,1][0,1][0,1] → a Finset (ℝ → ℝ) of lifts extended to R\mathbb RR by clamping the score to [0,1][0,1][0,1]; reason: functions that agree on [0,1][0,1][0,1] have the same extension, so the cardinality is unchanged.
Preamble
import Mathlib
import Definitions.Def_CalibratedMA_Randomized_Rounding
Formal statement
namespace CalibratedMA.Randomized

/-- Section 3.4, the counting bound and Eq. (5) (p. 17): with `ρ = 1/K`, `K_ρ = |Λ_ρ| = K + 1`,
`|A_{M,ρ}| ≤ Σ_{L=1}^M (K_ρ choose L+1) 2^{L+1} ≤ (M + 1)(2K_ρ)^{M+1}` and `log |A_{M,ρ}| ≤ C M log(e/ρ)` for a universal `C`. -/
theorem eq_5 :
    ∃ C : ℝ, 0 < C ∧ ∀ K M : ℕ, 1 ≤ K → 1 ≤ M →
      (liftedClass K M).card ≤
        ∑ L ∈ Finset.Icc 1 M, Nat.choose (K + 1) (L + 1) * 2 ^ (L + 1) ∧
      (∑ L ∈ Finset.Icc 1 M, Nat.choose (K + 1) (L + 1) * 2 ^ (L + 1)) ≤
        (M + 1) * (2 * (K + 1)) ^ (M + 1) ∧
      Real.log (liftedClass K M).card ≤ C * M * Real.log (Real.exp 1 / (1 / (K : ℝ))) := by sorry

end CalibratedMA.Randomized
Source
Jiuyao Lu, Georgy Noarov, Aaron Roth, Calibrated Multiaccuracy: Deterministic and Randomized Sample Complexity, STOC 2027 submission #900 (October 2026), p. 17, Section 3.4, counting display and Eq. (5)
Human review
  • Endorsed by Shuze Chen · Nov 10, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Nov 10, 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