Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge 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.

Finite Fourier inversion for the unitary transform on Z/NZ

Statement

Let N≥1 and f∈CZ/N. Then for every x∈Z/NZ

f(x)=N−1/2∑k=0N−1(FNf)(k) e2πikx/N,

the right-hand side being independent of the chosen integer representative of the class x. Consequently FN is bijective with inverse the positive-sign transform

(GNg)(x):=N−1/2∑k=0N−1g(k) e2πikx/N,g∈CZ/N, x∈Z/NZ,

and there is no convergence, regularity or support hypothesis anywhere.

Facts & Assumptions

Given: A natural number N≥1, a function f∈CZ/N, classes x,y∈Z/NZ, and integers j,ℓ.

[F1]

(FNh)(k)=N−1/2∑x=0N−1h([x]N)e−2πikx/N for every h∈CZ/N and integer k (The unitary discrete Fourier transform on Z/NZ).

[F2]

For all integers a,b, ∑q=0N−1e2πi(a−b)q/N=N if a≡b(modN) and 0 otherwise (Orthogonality of the characters x↦e2πikx/N on Z/NZ).

[F3]

Finite sums over Z/NZ: computed from any enumeration, invariant under reindexing along a bijection, additive over disjoint unions, Fubini, and scalars move through them (A finite sum in a commutative monoid indexed by an arbitrary finite set, Finite commutative-monoid sums are invariant under bijective reindexing, split over disjoint unions, and satisfy the finite Fubini rule); the classes [0]N,…,[N−1]N enumerate Z/NZ without repetition (For n≥1, every class in Z/n has one representative r with 0≤r<n, so ∣Z/n∣=n; while Z/0 is in bijection with Z).

[L2]

Classes: [u]N=[v]N exactly when u≡v(modN) (The congruence class [a]n and the quotient set Z/n), and −[u]N=[−u]N with −(−x)=x in the abelian group Z/NZ (For every natural n, (Z/n,+) is an abelian group, multiplication is a commutative monoid operation, and both distributive laws hold).

[L3]

Rational powers: N−1/2N−1/2=N−1 and N−1N=1 (Rational powers ar of a positive base, Laws of rational exponents).

Proof

technique · direct
1.1F1F3L1L3

Fix a class x and let x0∈{0,…,N−1} be its unique standard representative from [F3]. Substituting [F1] and using the product rule for exponentials [L1], the candidate right-hand side evaluated at x0 equals N−1/2∑k=0N−1(N−1/2∑y=0N−1f([y]N)e−2πiky/N)e2πikx0/N=N−1∑k=0N−1∑y=0N−1f([y]N)e2πi(x0−y)k/N, where N−1/2N−1/2=N−1 by [L3] and the interchange of the two finite sums is the Fubini rule [F3].

1.2F2L2

The inner sum over k is ∑k=0N−1e2πi(x0−y)k/N=N when [x]=[y]N and 0 otherwise: apply [F2] with a=x0, b=y and summation index k; the condition x0≡y(modN) is exactly [x]=[y]N by [L2].

1.3F3

Collapsing a sum supported at one class: for any c:Z/NZ→C and class a, ∑y∈Z/Nc(y)⋅(N if y=a, 0 otherwise)=N c(a). Split the finite index set into {a} and its complement by [F3]; the complement contributes 0 because every term there has the factor 0C, and the single term over {a} is the listed value.

1.4L1L2

The right-hand side depends only on the class of x: replacing its standard representative x0 by any representative x0+jN changes the exponent 2πikx0/N to 2πikx0/N+2πikj, and e2πikj=1 by [L1] because 2πikj∈2πiZ; so every summand, and hence the whole sum, is unchanged.

2.1F3L4step 1.1step 1.2step 1.3step 1.4

Therefore, for every class x, the right-hand side of the statement equals N−1∑y=0N−1f([y]N)(N if [y]N=[x], 0 otherwise)=N−1⋅N f(x)=f(x), by steps 1.1, 1.2 and 1.3, the list [0]N,…,[N−1]N containing one representative of every class by [F3]. This proves the inversion formula, and by step 1.4 the formula is a statement about the class x.

3.1F1F2F3L1L2L3L4step 1.1step 1.2step 1.3step 2.1

The transform GN of the statement is well defined by the same periodicity argument as step 1.4 (with the sign of the exponent reversed, which does not affect e2πik=1), and the computation of steps 1.1-2.1 with (f,FN,e−2πi⋅/N) replaced throughout by (g,GN,e+2πi⋅/N) gives FN(GNg)(k)=N−1∑yg([y]N)(N if [y]=[k],0 otherwise)=g(k) for every class k; that is, FN∘GN=id, while step 2.1 with g=FNf is GN∘FN=id.

4.1L5step 3.1∎

Since FN∘GN=id and GN∘FN=id, the transform FN has a two-sided inverse, namely GN; by [L5] FN is bijective and GN=FN−1, which is the statement.

Remarks

  • The exchange of signs is not a second theorem. The two compositions in step 3.1 are the same finite computation with the roles of x and k exchanged: both reduce to the orthogonality sum of [F2]. Both are verified because [L5] is stated for a two-sided inverse; no dimension argument and no countability or convergence argument is used.

  • Nothing here is a limit. All sums are finite, and the only scalar identity used beyond the orthogonality lemma is N−1N=1. In particular the inversion formula is exact for every function in CZ/N, including the zero function, and at N=1 it reads f([0])=1−1/2(F1f)(0)⋅1=f([0]), since F1=id.

Depends on

Used by

Dependency tree · two levels

76 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