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 thresholds rule out an r-independent VC covering bound
Statement
For each positive integer , let be the natural number obtained by multiplying by itself times. Equip with the uniform probability . For , define the threshold function , with . Put .
These are nested threshold indicators. Their VC-subgraph index is , while for and ,
Thus a purported entropy bound whose right side has no dependence on is false even when the VC index is fixed.
Facts & Assumptions
Given: A positive integer , the finite uniform probability space, and the threshold class in the Counterexample.
A class of functions is VC-subgraph when its measurable strict subgraphs have finite VC index (VC-subgraph classes, envelopes, and the first-unshattered index); for set classes, the index is one plus the VC dimension (Shattering, VC dimension, and first-unshattered index for set classes).
The distance is the th-root of the integral of the absolute difference to power , and covers use centers in the class (Covering, packing, bracketing numbers, and metric entropy).
Positive-base real powers satisfy , are defined by , and agree with rational powers at rational exponents (Real powers for positive bases, with the zero-base positive-exponent convention, The exponent, product, quotient, and iterated-power laws for positive real bases and real exponents, Rational powers of a positive base, Laws of rational exponents, The exponential definition of real powers agrees with the existing rational powers, Order, continuity, range, and the product, quotient, and reciprocal laws for the natural logarithm).
A finite uniform law assigns mass to each of its points and is a probability measure by finite additivity (Probability measures and probability spaces).
The natural powers grow beyond every fixed real bound: by the binomial theorem, and the Archimedean property supplies arbitrarily large (Integer powers , The binomial theorem in : , Every complete ordered field is Archimedean).
Proof
Given: The data in the Counterexample.
The real-power definition gives . Its exponent law gives for every natural , so induction identifies the real power with the natural product defining . By [F3],
The subgraphs are nested in increasing . A point belongs to and not to , so a singleton is shattered. No two-point set is shattered by a nested family, since shattering would require both opposite singleton traces, which nested sets cannot realize. Therefore the VC-subgraph dimension is exactly one and its index is .
For , the functions differ on exactly atoms. By [F2] and [F4],
Since , a closed ball of radius centered at a class member cannot contain two distinct class members: the triangle inequality would make their distance at most . Thus every cover needs at least centers; taking each member as its own center gives a cover with . This proves the exact covering number.
Fix any proposed finite right side that depends only on , the VC index, and constants independent of . At and it is a fixed finite number. By [F5], choose so large that exceeds that number. Step 3.1 then contradicts the proposed bound, so this family rules out a VC covering bound whose right-hand side is independent of .
Depends on
- For every $\varepsilon > 0$ in a complete ordered field there is a natural $n \ge 1$ with $1/n < \varepsilon$
- Covering, packing, bracketing numbers, and metric entropy
- The set $[A]^{k}$ of $k$-element subsets and the binomial coefficient $\binom{n}{k} := \lvert [n]^{k}\rvert$
- The canonical natural $\iota(n) = n \cdot 1_F$ of a field
- Integer powers $a^m$
- The natural numbers $\mathbb{N}$ (von Neumann)
- Probability measures and probability spaces
- Rational powers $a^r$ of a positive base
- Real powers for positive bases, with the zero-base positive-exponent convention
- Shattering, VC dimension, and first-unshattered index for set classes
- VC-subgraph classes, envelopes, and the first-unshattered index
- Laws of rational exponents
- The exponential definition of real powers agrees with the existing rational powers
- The exponent, product, quotient, and iterated-power laws for positive real bases and real exponents
- Every complete ordered field is Archimedean
- The binomial theorem in $\mathbb{R}$: $(x+y)^{n} = \sum_{k<n+1} \iota\!\binom{n}{k}\, x^{k} y^{\,n-k}$
- The principle of mathematical induction
- Order, continuity, range, and the product, quotient, and reciprocal laws for the natural logarithm
Used by
Nothing in the library uses this result yet.
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
- Jon A. Wellner, Empirical Processes: Theory and Applications, Chapter 1, Theorem 8.1 (standard reference, not scraped)
- David Pollard, Probability Tools, Tricks, and Miracles, 2026 manuscript, Section 14.5 (standard reference, not scraped)