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 positive semidefinite kernel has a canonical reproducing kernel Hilbert space under countable choice
Statement
Assume the Axiom of Countable Choice. For every set and every real positive semidefinite kernel on , the canonical construction from finitely supported coefficients, quotient by zero seminorm, and norm completion defines a real reproducing kernel Hilbert space . For every , its kernel section belongs to and satisfies for all . The choice assumption here is countable choice; this statement does not assert the same construction in ZF without that assumption.
Statement references
Hastie, Tibshirani, and Friedman, The Elements of Statistical Learning, §5.8
Rasmussen and Williams, Gaussian Processes for Machine Learning, §6.1
Positive semidefinite kernels and reproducing kernel Hilbert spaces
Every metric space has a completion, constructed as the equivalence classes of its Cauchy sequences
The metric completion of a normed space carries a unique compatible Banach-space structure
The norm completion of an inner-product space is a Hilbert space
Facts & Assumptions
Given: The Axiom of Countable Choice, a set , and a real positive semidefinite kernel on .
The Axiom of Countable Choice says every countable sequence of nonempty sets has a choice function (The Axiom of Countable Choice ()).
A real positive semidefinite kernel is symmetric and every finite Gram quadratic form is nonnegative (Positive semidefinite kernels and reproducing kernel Hilbert spaces).
The completion used here is obtained by taking the metric completion (A completion of a metric space: a complete metric space together with an isometric embedding onto a dense subspace, Every metric space has a completion, constructed as the equivalence classes of its Cauchy sequences), equipping it with its unique compatible Banach structure under countable choice (The metric completion of a normed space carries a unique compatible Banach-space structure), viewing it as a norm completion (Completion of a normed space), and extending the real inner product uniquely under countable choice (The norm completion of an inner-product space is a Hilbert space).
In a real inner-product space, (Cauchy–Schwarz: , with equality exactly for linearly dependent vectors).
Proof
Let be the finitely supported functions on and define , where each sum is over the finite supports. Symmetry and bilinearity follow from symmetry of and finite-sum algebra. For any , listing its finite support gives by [F1].
For and , . If , nonnegativity of this quadratic for all gives ; if , its linear coefficient must vanish, so .
It follows that every satisfies for every . Hence is a subspace and the form descends to a well-defined positive definite inner product on : changing a representative by an element of does not change any pairing, and a quotient class has zero squared norm exactly when it is the zero class.
The induced inner product makes a real inner-product space. By [F2] and [A1], its norm metric has a metric completion with compatible Banach structure, hence a norm completion; the inner product extends uniquely and makes the completion a real Hilbert space.
For each , the image of satisfies . Thus the finite kernel sections are elements of , and their span is dense because the image of is dense in its completion.
For , the quotient pairing gives , independently of the representative by [step 3.1]. Given , density in [step 4.1] provides an approximating sequence . By [F3], the real sequence is Cauchy, and if is another such sequence then , so its limit is independent of the approximants. The extended inner product also satisfies [F3], hence this limit equals .
Define the associated function of by . Applying the extended pairing to gives , so is exactly the section and for every .
If for every , then is orthogonal to every section and therefore to their dense span from [step 5.1]. Continuity of the inner product gives , so positive definiteness gives . Thus the associated-function map is injective and identifies with a Hilbert space of functions having reproducing kernel .
Depends on
- Positive semidefinite kernels and reproducing kernel Hilbert spaces
- Completion of a normed space
- A completion of a metric space: a complete metric space together with an isometric embedding onto a dense subspace
- Hilbert space
- The Axiom of Countable Choice ($\mathrm{AC}_\omega$)
- Every metric space has a completion, constructed as the equivalence classes of its Cauchy sequences
- The metric completion of a normed space carries a unique compatible Banach-space structure
- The norm completion of an inner-product space is a Hilbert space
- Cauchy–Schwarz: $|\langle u,v\rangle|\leq\lVert u\rVert\lVert v\rVert$, with equality exactly for linearly dependent vectors
Used by
- Gaussian RKHS power function decays exponentially with inverse fill distance Lemma
- Noiseless Gaussian RKHS interpolation error under uniform sampling Theorem
Cited to discharge well-definedness by Positive semidefinite kernels and reproducing kernel Hilbert spaces.
Dependency tree · two levels
51 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
- Hastie, Tibshirani, and Friedman, The Elements of Statistical Learning, 2nd ed., corrected 12th printing (standard reference, not scraped)
- Rasmussen and Williams, Gaussian Processes for Machine Learning (standard reference, not scraped)