Alphabeta Math
ExampleConstruction: Literature-sourcedVerification: AI-adaptedPipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Finite lacunary exponential sums belong to BMO

Example

Let 0<a≤b<∞, and let (ξn)n∈Z be nonzero real frequencies satisfying a2n≤∣ξn∣≤b2n for every n. Let (cn)n∈Z have finite support and put f(x):=∑n∈Zcne2πiξnx. Then ∥f∥BMO(R)≤Ca,b,ϕ(∑n∣cn∣2)1/2. More precisely, fix a nonnegative ϕ∈Cc∞(R) with ϕ≥1 on [−12,12]. For an interval I of length L and centre xI, set ϕI(x):=ϕ((x−xI)/L). There is a constant cI∈C such that ∫RϕI(x)∣f(x)−cI∣2 dx≤Ca,b,ϕL∑n∣cn∣2.

Facts & Assumptions

Given: The finite exponential sum f, the frequency bounds above, a nonnegative compactly supported smooth bump ϕ with ϕ≥1 on [−12,12], and a nondegenerate interval I.

[L1]

For every locally integrable g and every constant c, ∣I∣−1∫I∣g−gI∣≤2∣I∣−1∫I∣g−c∣, and the BMO seminorm is the supremum of the mean oscillations over intervals in dimension one (BMO seminorm and the quotient by constants).

[L2]

Since ϕ∈Cc∞, both ∫ϕ(y) dy and ∫ϕ(y)y2 dy are finite. For every integer M≥1, integration by parts M times gives ∣∫Rϕ(y)e2πiζy dy∣≤CM,ϕ(1+∣ζ∣)−M. For ∣ζ∣≤1 this follows instead from ∥ϕ∥1; combining the two estimates gives the displayed bound.

Verification

technique · direct low- and high-frequency split

Fix I, write L=∣I∣, and let n0 be the least integer with 2n0≥L−1. Thus 1≤L2n0<2. Split f=f<+f≥ at n0 and choose cI:=f<(xI)=∑n<n0cne2πiξnxI.

1.1L2algebra

Low frequencies. For every n<n0, ∣e2πiξnx−e2πiξnxI∣≤2π∣ξn∣∣x−xI∣. After x=xI+Ly, (∫ϕI(x)∣e2πiξnx−e2πiξnxI∣2dx)1/2≤Cϕ∣ξn∣L3/2. The triangle inequality in L2(ϕIdx), Cauchy--Schwarz, and the upper frequency bound therefore give ∥f<−cI∥L2(ϕIdx)≤CϕL3/2∑n<n0∣cn∣∣ξn∣≤Cb,ϕL3/2(∑n∣cn∣2)1/2(∑n<n04n)1/2≤Cb,ϕL1/2(∑n∣cn∣2)1/2, because 2n0<2L−1.

1.2L2algebra

High frequencies. Define Gnm:=L−1∫RϕI(x)e2πi(ξn−ξm)x dx(n,m≥n0). The change of variables x=xI+Ly and [L2] give, for every fixed integer M≥1, ∣Gnm∣≤CM,ϕ(1+L∣ξn−ξm∣)−M. Choose an integer K≥1 such that b2−K≤a/2. If ∣n−m∣>K, the frequency comparability implies ∣ξn−ξm∣≥∣∣ξn∣−∣ξm∣∣≥(a/2)2max⁡(n,m). For each fixed n≥n0, the terms with ∣m−n∣≤K contribute at most a constant to ∑m≥n0∣Gnm∣. For m>n+K their sum is bounded by C∑r≥1(1+L2n+r)−M≤C∑r≥12−Mr<∞, using L2n≥1. For n0≤m<n−K, there are at most n−n0 terms, each bounded by C(1+L2n)−M; their total is bounded uniformly because L2n≥2n−n0 and sup⁡r≥0r(1+2r)−M<∞. Hence the absolute row sums of (Gnm)n,m≥n0 are uniformly bounded. Since ∣Gnm∣=∣Gmn∣, the inequality 2∣cn∣∣cm∣≤∣cn∣2+∣cm∣2 yields ∫ϕI∣f≥∣2≤L∑n,m≥n0∣cn∣∣cm∣∣Gnm∣≤Ca,b,ϕL∑n∣cn∣2.

2.1step 1.1step 1.2L1algebra∎

Combining the two pieces with ∣f−cI∣2≤2∣f<−cI∣2+2∣f≥∣2 gives ∫RϕI∣f−cI∣2≤Ca,b,ϕL∑n∣cn∣2. Since ϕI≥1 on I, Cauchy--Schwarz and [L1] imply L−1∫I∣f−fI∣≤2L−1∫I∣f−cI∣≤2L−1/2(∫I∣f−cI∣2)1/2≤Ca,b,ϕ(∑n∣cn∣2)1/2. Taking the supremum over all nondegenerate intervals proves the BMO bound. The proof uses finite sums only and no choice principle.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

5 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources