Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-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.

Expected-norm estimate from the corrected moment comparison

PDF p. 34: Proof of Proposition 6.3, final displays

Statement

Assume the Axiom of Choice. Let n≥1 and ℓ≥1 be integers and p∈[0,1]. Let A be the adjacency matrix of the loop-free Erdős–Rényi graph on n vertices: distinct unordered edges occur independently with probability p. Put ΔA=A−pJ, where J is the all-ones matrix, and use the Euclidean operator norm. Set q=2p(1−p). Then

E∥ΔA∥≤p+q1/(2ℓ)(2n+22ℓ)elog⁡n/(2ℓ).

For α>0 and real x≥3, define f(α,x)=22e1/(2α)(1+2αlog⁡x/x). This function is strictly decreasing in x. If ℓ=αlog⁡n is a positive integer, the moment estimate becomes

E∥ΔA∥≤p+q1/(2ℓ)2n e1/(2α)(1+2αlog⁡n/n).

The article’s stronger displayed estimate is E∥ΔA∥≤q(2n+22ℓ)elog⁡n/(2ℓ); it does not hold uniformly for p∈(0,1). In particular, the substitution ℓ=αlog⁡n is available to the integer-moment argument only when that quantity is an integer.

Facts & Assumptions

[F1]

For an independent copy A′, put W=A−A′. Corrected symmetrization gives E∥ΔA∥≤p+E∥W∥ and E∥W∥≤(ETr⁡W2ℓ)1/(2ℓ). Here Tr⁡ is the sum of diagonal entries. Symmetrization and moments of the adjacency difference

[F2]

If G is symmetric with independent standard normal entries on and above the diagonal, reflected below it, the corrected comparison is ETr⁡W2ℓ≤q ETr⁡G2ℓ. Trace-moment comparison with Gaussian matrices, Gaussian comparison matrix in the moment argument

[F3]

For that same Gaussian convention, (ETr⁡G2ℓ)1/(2ℓ)≤n1/(2ℓ)(2n+22ℓ). The Gaussian construction assumes Choice. Gaussian trace/norm estimate attributed to Slepian comparison, The Axiom of Choice

[F4]

The derivative of log⁡x is 1/x on (0,∞), and the mean-value theorem gives strict decrease on an interval when the derivative is everywhere negative. The natural logarithm has derivative 1/x and equals the integral from 1 to x of 1/t, Mean value theorem on a closed real interval

[F5]

If real functions u,v are differentiable at x and v(x)≠0, then (u/v)′(x)=(u′(x)v(x)−u(x)v′(x))/v(x)2. Sums, scalar multiples, products and quotients: (f+g)'(c) = f'(c) + g'(c), (α f)'(c) = α f'(c), (fg)'(c) = f'(c)g(c) + f(c)g'(c), and (f/g)'(c) = (f'(c)g(c) - f(c)g'(c))/g(c)² when g(c) ≠ 0

[F6]

For t≥0, log⁡(1+t)≥2t/(2+t), with equality precisely at t=0 (Angular order and elementary inverse-trigonometric bounds).

Proof

Given: The parameters, graph and Choice assumption in the Statement.

Goal: Prove the corrected estimate, the coefficient’s monotonicity, and the failure of the stronger displayed estimate.

1.1F1F2F3givenalgebra

Apply [F1] and [F2], then take nonnegative 2ℓ-th roots and use [F3]. This yields E∥ΔA∥≤p+q1/(2ℓ)n1/(2ℓ)(2n+22ℓ). Since n1/(2ℓ)=elog⁡n/(2ℓ), this is the first estimate, including q=0.

1.2F4F5F6givenalgebra

The identity function has derivative one by its difference quotient. Apply the quotient rule [F5] to u(x)=log⁡x and v(x)=x on (0,∞): [F4] supplies u′(x)=1/x, and v(x)>0. The strict equality case in [F6], applied at t=2>0, gives log⁡3>1. The positive logarithm derivative in [F4] and its mean-value theorem give log⁡x≥log⁡3>1 for x≥3. Thus the derivative of log⁡x/x is (1−log⁡x)/x2<0 there. By [F4], log⁡x/x is strictly decreasing on [3,∞). The square-root function is strictly increasing on positive arguments: if 0<a<b and a≥b, squaring would contradict a<b. The remaining factors in f are positive and independent of x, so f(α,x) strictly decreases on [3,∞).

2.1step 1.1givenalgebra

When ℓ=αlog⁡n, substitute this equality into step 1.1 and use 2n+22ℓ=2n(1+2ℓ/n). This gives the stated specialized estimate with factor 2n. The moment comparison requires integer ℓ, so the substitution has that stated restriction.

3.1givenalgebra∎

For fixed n≥2 and integer ℓ≥1, every diagonal entry of ΔA is −p; applying it to a standard unit vector gives ∥ΔA∥≥p. As p↑1, the article’s proposed right-hand side 2p(1−p)(2n+22ℓ)elog⁡n/(2ℓ) tends to zero, while E∥ΔA∥≥p→1. Therefore the proposed inequality fails for all sufficiently large p<1, for every fixed n and ℓ in this range.

Source notes

The original article displays the stronger estimate and the coefficient f at PDF p. 34. Its moment comparison uses qℓ where the completed comparison supplies q, and its symmetrization omits the p term. The article citation is attribution only. The completed suppliers, rather than that display, provide the premises used above.

Depends on

Used by · 2

Dependency tree · 2 levels

Each result is drawn at its shortest distance from this item. Arrows run from prerequisites to the items that use them; this item has a heavier outline. External source nodes link directly to the cited text. The view shows up to two dependency steps; open a prerequisite to continue.

Sources