Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 8.2 — with η\etaη anomalies, optRAND(F,η)≥12 opt(F,0)+η\mathrm{opt}_{\mathrm{RAND}}(F,\eta) \ge \tfrac12\,\mathrm{opt}(F,0) + \etaoptRAND​(F,η)≥21​opt(F,0)+η

Proved
WeightedMajority.RandAnomalies.theorem_8_2

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

anomalieslower-boundmistake-boundonline-learningp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1randomized-algorithmsweighted-majority

Let FFF be a pool of {0,1}\{0,1\}{0,1}-valued functions on an arbitrary domain XXX, let η≥0\eta \ge 0η≥0 be an integer, let opt(F,0)\mathrm{opt}(F,0)opt(F,0) be the optimal worst-case number of mistakes of a deterministic learner on sequences consistent with some member of FFF, and let optRAND(F,η)\mathrm{opt}_{\mathrm{RAND}}(F,\eta)optRAND​(F,η) be the optimal worst-case expected number of mistakes of a randomized learner on sequences with at most η\etaη anomalies with respect to FFF.

Theorem 8.2. If FFF contains at least two functions, then

optRAND(F,η) ≥ 12 opt(F,0)+η.\mathrm{opt}_{\mathrm{RAND}}(F,\eta) \ \ge\ \tfrac12\,\mathrm{opt}(F,0) + \eta .optRAND​(F,η) ≥ 21​opt(F,0)+η.

Thus every randomized learner has worst-case expected mistakes at least 12opt(F,0)+η\tfrac12\mathrm{opt}(F,0)+\eta21​opt(F,0)+η over sequences fixed in advance with at most η\etaη anomalies. When opt(F,0)=∞\mathrm{opt}(F,0) = \inftyopt(F,0)=∞, these expected mistake counts are unbounded. It is the randomized counterpart of Theorem 8.1 (opt(F,η)≥opt(F,0)+2η\mathrm{opt}(F,\eta) \ge \mathrm{opt}(F,0) + 2\etaopt(F,η)≥opt(F,0)+2η), and shows that the coefficient 111 of η\etaη in the expected mistake bound of the randomized Weighted Majority algorithm WMR cannot be improved by any randomized algorithm.

Formalization Note opt(F,0)∈N∪{∞}\mathrm{opt}(F,0) \in \mathbb N\cup\{\infty\}opt(F,0)∈N∪{∞} is cast to [0,∞][0,\infty][0,∞], where 12⋅∞+η=∞\tfrac12\cdot\infty + \eta = \infty21​⋅∞+η=∞. "∣F∣>1|F| > 1∣F∣>1" is F.Nontrivial (two distinct members), which also covers infinite pools.

Preamble
import Definitions.Def_WeightedMajority_RandAnomalies_OptRand

open scoped ENNReal
Formal statement
namespace WeightedMajority.RandAnomalies

/-- **Theorem 8.2** (Littlestone–Warmuth 1994, p. 252). For all target classes `F` and all
`η ≥ 0`, if `|F| > 1` then `opt_RAND(F, η) ≥ ½ opt(F, 0) + η`. -/
theorem theorem_8_2 {X : Type*} (F : Set (X → Bool)) (hF : F.Nontrivial) (η : ℕ) :
    ((WeightedMajority.Anomalies.opt F 0 : ℕ∞) : ℝ≥0∞) / 2 + η ≤ optRand F η := by sorry

end WeightedMajority.RandAnomalies
Source
Littlestone, Warmuth, The Weighted Majority Algorithm, Inform. and Comput. 108 (1994), p. 252, Theorem 8.2
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 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