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.
Measurable minimum-norm geometric-median selection
Statement
Fix a separable real Hilbert space and an integer . The rule in Geometric median of block means defines a single-valued Borel map : for every tuple the function has a nonempty closed convex minimizer set, and that set has a unique point of minimum norm. The choice-free Borel construction below does not use weak compactness or reflexivity. Countable Choice is used only for the published Hilbert projection supplier that identifies the minimum-norm point.
Facts & Assumptions
Given: A separable real Hilbert space , an integer , and a tuple .
The geometric-median objective is ; the algorithm selects the minimum-norm point among its minimizers (Geometric median of block means).
In a Hilbert space, a nonempty closed convex set has a unique nearest point to under Countable Choice (Projection onto a nonempty closed convex set, The Axiom of Countable Choice ()).
A closed bounded subset of is compact for finite (Heine-Borel in : with the Euclidean metric a subset of is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line).
Cauchy–Schwarz implies in a Hilbert space (Cauchy–Schwarz: , with equality exactly for dependent pairs).
A continuous real-valued function on a nonempty compact metric space attains a minimum and maximum (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value).
In a compact metric space, every family of closed sets with the finite-intersection property has nonempty intersection (A metric space is compact if and only if every family of closed subsets with the finite intersection property has nonempty intersection).
Borel sets form a sigma-algebra, a map is Borel when inverse images of Borel sets are Borel, countable infima of Borel real maps are Borel, and pointwise limits of Borel real maps are Borel (The Borel sigma-algebra of a topological space, A measurable function between measurable spaces, Sequential suprema, infima, limsup, liminf, and pointwise limits of measurable functions are measurable).
Separability means there is an at-most-countable dense subset; since is nonempty, A nonempty set is at most countable iff it is a surjective image of enumerates it by a sequence (Separability: the existence of an at most countable dense subset).
Proof
Given: A separable real Hilbert space , an integer , and a tuple .
Let . In the given center order, Gram–Schmidt replaces by using the previously retained orthonormal vectors; if it is skipped, and otherwise is appended. Induction on the finite list shows the retained vectors are orthonormal and span . For , if converges in to , then [F4] gives for each ; the finite sums converge to , so is closed. If , all centers are zero and has the unique minimizer . For , set ; its residual is orthogonal to . Pythagoras gives , so whenever . Thus every minimizer lies in . The function is continuous and coercive because ; its sublevel set is nonempty, closed and bounded in the finite-dimensional coordinate space , hence compact by [F3]. By [F5], attains a minimum there; the full minimizer set is nonempty, closed, bounded and compact, and it is convex because is convex.
By [F2], has a unique point nearest to , its unique minimum-norm point. If , the selection is the constant zero map; otherwise fix a unit vector and a countable dense sequence in , whose existence is given by [F8]. The countable set is dense in : for each , density gives , and [F4] gives . For , is continuous, coercive and strictly convex. Projection to cannot increase and strictly decreases the squared norm off , so every minimizer lies in ; if the single candidate is . For , a closed bounded sublevel set in is compact by [F3], so [F5] gives a unique minimizer . By continuity and density, . Each map is continuous, so [F7] makes their countable infimum Borel. For , set . The countable dense set gives so is Borel; density ensures at least one such for each . The least such is a Borel, countably-valued map because its level set at is , and the selected point map is Borel since each open-set preimage is the union of its Borel level sets whose values lie in that open set. The selected point is within of : convexity of and , applied at the minimizer and followed by , give . Hence the selected maps converge pointwise to . Since they are countably valued with Borel level sets, their distances to each are Borel and converge to ; [F7] makes these limits Borel. Balls with centers and positive rational radii form a countable basis of the separable metric space : every open neighborhood of a point contains such a ball around a sufficiently close dense center. Thus inverse images of open sets are countable unions of Borel inverse images of these balls, proving is Borel.
For every , comparison with in the minimization of gives and . If then both points are . Otherwise all lie in the compact ball by [F3]. The closed tail sets inside this ball are nonempty, nested and have the finite-intersection property, so [F6] gives . For every , continuity of and the tail bounds imply , while ; letting shows is a minimum-norm minimizer, hence . If the sequence did not converge to , some would leave arbitrarily late terms at distance at least ; the closed tail closures of those terms are nonempty nested subsets of the same compact ball, so [F6] gives a point in their intersection. It lies in every , hence equals , but its distance from is at least , a contradiction. Thus . By [F7] and the separable-metric basis argument of step 1.2, the pointwise limit of the Borel maps is Borel.
The minimizer set is nonempty, closed and convex by step 1.1, its minimum-norm point is unique by step 1.2, and step 2.1 proves the selected map is Borel. The only choice principle used is Countable Choice in [F2]; the dense sequence is supplied by separability and all subsequent approximations use least-index rules.
Depends on
- Geometric median of block means
- Hilbert space
- The Axiom of Countable Choice ($\mathrm{AC}_\omega$)
- Separability: the existence of an at most countable dense subset
- A nonempty set is at most countable iff it is a surjective image of $\mathbb{N}$
- Projection onto a nonempty closed convex set
- Heine-Borel in $\mathbb{R}^n$: with the Euclidean metric a subset of $\mathbb{R}^n$ is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line
- Cauchy–Schwarz: $|\langle x,y\rangle|\le\|x\|\,\|y\|$, with equality exactly for dependent pairs
- A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value
- A metric space is compact if and only if every family of closed subsets with the finite intersection property has nonempty intersection
- The Borel sigma-algebra of a topological space
- A measurable function between measurable spaces
- Sequential suprema, infima, limsup, liminf, and pointwise limits of measurable functions are measurable
Used by
- Adaptive truncated covariance estimator Algorithm
- A polynomial-tail law with finite variance Example
- One corrupted block is rejected by the geometric median Example
- Exact block budget for the covariance geometric-median pilot Lemma
- The adaptive truncated covariance procedure is well-defined on its declared domain Lemma
- Finite-variance Hilbert mean bound for geometric medians of means Theorem
Cited to discharge well-definedness by Geometric median of block means.
Dependency tree · two levels
83 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
- Stanislav Minsker, Geometric Median and Robust Estimation in Banach Spaces, §2 and §4.1 (standard reference, not scraped)
- Gábor Lugosi, Mean Estimation in High Dimension, §3 (standard reference, not scraped)