Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-6.1-sol)
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.

Mollification rates for compactly supported Slobodeckij functions

Statement

Assume the Axiom of Countable Choice. Let d≥1, 0<θ<1, 1≤p<∞, let g∈Wθ,p(Rd) have compact support, and let gδ:=g∗ηδ be the mollification of g by a radial mollifier η with ∫η=1. Then gδ∈Cc∞(Rd) is supported in the δ-neighbourhood of supp⁡g; There are constants C1=C1(d,p,θ) and C2=C2(d,p,θ,η) such that (i) ∥g−gδ∥Lp(Rd)≤C1(d,p,θ) δθ[g]θ,p;(ii) ∥gδ∥W1,p(Rd)≤C2(d,p,θ,η)(∥g∥Lp(Rd)+δθ−1[g]θ,p).

Facts & Assumptions

Given: the Axiom of Countable Choice, d≥1, 0<θ<1, 1≤p<∞, a compactly supported g∈Wθ,p(Rd), a radial mollifier η with ∫η=1, and gδ=g∗ηδ. Write ω(t):=sup⁡∣z∣≤t∥τzg−g∥Lp(Rd) for t≥0.

[F1]

Slobodeckij seminorm as a translation integral. Because the diagonal is null and Tonelli's theorem together with the substitution z=y−x applies, [g]θ,pp=∫Rd∥τzg−g∥Lp(Rd)p∣z∣−d−pθ dz. (The Gagliardo--Slobodeckij space on Euclidean space, Tonelli and Fubini for the completed product, with only almost-everywhere section measurability, Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation, Translation of a function on Rn)

[F3]

The translation modulus is subadditive. ω is nondecreasing, and ω(s+t)≤ω(s)+ω(t) for all s,t≥0, because τzg−g=τz2(τz1g−g)+(τz2g−g) when z=z1+z2 and translations are isometries of Lp. (Translation of a function on Rn, Lebesgue outer measure, Lebesgue measurability and Lebesgue measure are unchanged by translation, The space Lp(μ) as the quotient by null functions)

[F4]

Minkowski's integral inequality. For measurable F on a product of sigma-finite measure spaces with ∫Y∥F(⋅,y)∥p dν(y)<∞, ∥∫YF(⋅,y) dν(y)∥p≤∫Y∥F(⋅,y)∥p dν(y). (Minkowski's integral inequality)

[F5]

Young's convolution inequality. ∥f∗h∥Lp≤∥f∥L1∥h∥Lp. (Young's convolution inequality under Countable Choice)

[F6]

Mollifying a locally integrable function. For g∈Lloc1 the convolution g∗ηδ is smooth with ∂α(g∗ηδ)=g∗(∂αηδ); an Lp function is locally integrable, and a compactly supported Lp function lies in L1. (Convolution with a mollifier is smooth, and derivatives pass under the integral sign, Holder's inequality for integrals, including the endpoint cases)

[F7]

Support of a convolution. For Borel representatives of f,h∈L1, supp⁡(f∗h)⊆supp⁡f+supp⁡h‾. (The support of a convolution lies in the closure of the support sumset)

[F8]

The radial mollifier. ηδ=δ−dη(⋅/δ) satisfies ∫ηδ=1, supp⁡ηδ⊆B(0,δ), and ∫Rd∂jηδ=0 for every j, since η is radial and its gradient is odd in each coordinate. (A radial mollifier family in Rn)

Proof

technique · Use subadditivity of the translation modulus to dominate its supremum by an average over a ball, compare that average to the Slobodeckij translation integral, and then estimate the approximation error and gradient by Minkowski's integral inequality
1.1F1F2F3

Fix t>0 and ∣h∣≤t. By [F3], for every z∈B(0,t) we have ∥τhg−g∥p≤∥τzg−g∥p+∥τh−zg−g∥p. Raising to the p-th power and averaging over z∈B(0,t) gives ∥τhg−g∥pp≤2p−1∣B(0,t)∣(∫B(0,t)∥τzg−g∥ppdz+∫B(0,t)∥τh−zg−g∥ppdz)≤2pvdtd∫∣w∣≤2t∥τwg−g∥ppdw, because both B(0,t) and h−B(0,t) lie in B(0,2t) and [F2] gives ∣B(0,t)∣=vdtd. Taking the supremum over ∣h∣≤t, then using [F1] and ∣w∣−d−pθ≥(2t)−d−pθ for 0<∣w∣≤2t, yields ω(t)p≤2pvdtd(2t)d+pθ[g]θ,pp. Thus ω(t)≤C1tθ[g]θ,p for a constant C1=C1(d,p,θ), which is the translation-modulus estimate needed below.

2.1F4F8step 1.1

Since ∫ηδ=1 and ηδ≥0 is supported in B(0,δ), g(x)−gδ(x)=∫ηδ(y)(g(x)−g(x−y))dy; taking Lp-norms and applying [F4] with Y=B(0,δ) gives ∥g−gδ∥p≤∫ηδ(y)∥g−τyg∥p dy≤ω(δ)≤C1δθ[g]θ,p, which is (i).

3.1F5F6F7F8step 1.1∎

By [F6], gδ is smooth and ∇gδ=g∗∇ηδ; [F6] also gives g∈L1 because g has compact support and lies in Lp. By [F8], ∫∂jηδ=0, so ∂jgδ(x)=∫(g(x−y)−g(x))∂jηδ(y) dy, and [F4] gives ∥∂jgδ∥p≤ω(δ)∫∣∂jηδ∣≤C2δθ−1[g]θ,p because ∫∣∇ηδ∣=δ−1∫∣∇η∣. Moreover ∥gδ∥p≤∥g∥p by [F5] with ∥ηδ∥1=1, so ∥gδ∥W1,p≤∥g∥p+∑j∥∂jgδ∥p, which is (ii) after enlarging the constant. Finally, gδ∈C∞ by [F6], and [F7] applied to the Borel representative of g and to ηδ gives supp⁡gδ⊆supp⁡g+B(0,δ)‾, the δ-neighbourhood of supp⁡g; this is compact because supp⁡g is compact, so gδ∈Cc∞(Rd).

Depends on

Used by

Dependency tree · two levels

93 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