Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6-luna)
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.

A Gaussian tail bound implies a sub-Gaussian MGF bound

Statement

Let X be a centered real random variable and let τ>0. If P(∣X∣>t)≤2e−t2/(2τ)(t>0), then X is sub-Gaussian with proxy 64τ in the MGF convention of Sub-Gaussian variables and vectors, and sub-exponential variables: EeλX≤e32τλ2(λ∈R).

Facts & Assumptions

Given: A centered real random variable X with the displayed two-sided tail bound and a constant τ>0.

[F1]

Layer cake gives ET=∫0∞P(T>t) dt for nonnegative T (Layer-cake formulas for random variables).

[F2]

A product of two probability laws exists and has the rectangle formula (For sigma-finite factors, the product measure exists, has the rectangle formula, is sigma-finite, and is unique).

[F3]

Expectations of products of independent real random variables factor (Expectations factor over finite products of independent random variables).

[F4]

Jensen's inequality gives Ee−λX≥e−λEX=1 (Jensen's inequality for expectation).

[F5]

Increasing nonnegative functions pass to the limit under expectation (Monotone convergence for the integral).

[F6]

A centered variable is sub-Gaussian with proxy v when EeλX≤evλ2/2 for all real λ (Sub-Gaussian variables and vectors, and sub-exponential variables).

Proof

1.1F1F2givenalgebra

The tail bound and [F1] imply Eea∣X∣=1+a∫0∞eatP(∣X∣>t) dt≤1+2a∫0∞eat−t2/(2τ) dt<∞ for every a>0, since completing the square gives at−t2/(2τ)=−(t−aτ)2/(2τ)+a2τ/2. Thus all MGFs below are finite.

1.2F2F3F4givenalgebra

Let X1,X2 be the coordinate variables on the product of two copies of the law of X; [F2] gives this probability space, and its rectangle formula makes them independent with law equal to that of X. Set D=X1−X2. Independence [F3] gives EeλD=(EeλX)(Ee−λX). By [F4], the second factor is at least 1, so EeλX≤EeλD. The product is unchanged when λ is replaced by −λ, hence EeλD=Ecosh⁡(λD).

2.1F1step 1.2algebra

A union bound gives P(∣D∣>t)≤4e−t2/(8τ) because ∣D∣>t implies ∣X1∣>t/2 or ∣X2∣>t/2. For m≥1, [F1] and the substitution u=t2/(8τ) yield E∣D∣2m=2m∫0∞t2m−1P(∣D∣>t) dt≤4(8τ)mm!. The integral is evaluated using ∫0∞t2m−1e−t2/(8τ)dt=12(8τ)m(m−1)!.

3.1F5F6step 2.1step 1.2algebra∎

By [F5], the nonnegative power series for cosh⁡ may be integrated term by term. With x=8τλ2, step 2.1 and (2m)!≥(m!)2 give Ecosh⁡(λD)≤1+4∑m≥1xmm!=4ex−3≤e4x. For the last inequality, f(x)=e4x−4ex+3 satisfies f(0)=0 and f′(x)=4(e4x−ex)≥0 for x≥0. Therefore EeλX≤e32τλ2 for every λ, which is the proxy-64τ assertion by [F6].

Depends on

Used by

Dependency tree · two levels

43 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