MANUSCRIPT ATTRIBUTION AND BIBLIOGRAPHY REVIEW
2026-10-18. This is a focused source/attribution review, not a full proof audit.
Reviewed source:
D:/LUMINARY ODDITIES LLC/AI Research/odd_prime_learner_prepublication_Riemann_A_20261018/manuscript/main.tex
SHA256: 63f1a1a2400046d0083e8c654a5730a6239a713f09560fce83bcea06153b05b1
No manuscript edits made. Bind any subsequent review to its new source hash.

DISPOSITION
No material bibliography or attribution blocker found in the inspected source.
The abstract/introduction/scope sections correctly credit known measurement
arithmetic and bootstrap architecture, distinguish pure versus mixed input, and
retain the qutrit draft within one result family. The exact operation model,
fixed-p versus growing-p resources, no minimal-copy assertion, and no worldwide
priority claim are appropriately visible. These editorial findings do not
certify the analytic inequalities or complete learner resource proof.

BIBLIOGRAPHY
The five central entries match current primary arXiv records reopened in this
review and the immediately preceding refresh:
- Qi / Xu / Feng / Li, Agnostic learning of qudit stabilizer states,
  2607.15559v1, 17 July 2026.
- Chen / Gong / Ye / Zhang, Stabilizer bootstrapping: A recipe for efficient
  agnostic tomography and magic estimation, 2408.06967v3, 5 December 2024.
- Allcock / Doriguello / Ivanyos / Santha, Reconquering Bell sampling on qudits:
  stabilizer learning and testing, quantum pseudorandomness bounds, and more,
  2510.06848v3, 25 August 2026.
- Hinsche / Eisert / Carrasco, The abelian state hidden subgroup problem:
  Learning stabilizer groups and beyond, 2505.15770v3, 27 May 2026.
  J. Carrasco is correct (Jose Carrasco). The paper begins its title with The.
- Noller / Tran / Gachechiladze / Kueng, An infinite hierarchy of multi-copy
  quantum learning tasks, 2510.08070v2, 1 September 2026.
  The source's LaTeX accent N\"oller and initials V. T. Tran are correct.

The Wootters/Fields entry matches publisher-indexed metadata: William K. Wootters
and Brian D. Fields, Optimal state-determination by mutually unbiased
measurements, Annals of Physics 191(2), 363-381, 1 May 1989,
DOI 10.1016/0003-4916(89)90322-9.
https://www.sciencedirect.com/science/article/pii/0003491689903229
The publisher metadata was available in search retrieval. Direct publisher-page
retrieval failed, and the Chrome page displayed a human-verification challenge;
it was not solved. Thus this is metadata verification, not a fresh full-text
read of Wootters/Fields. The manuscript uses it for standard MUB background,
while deriving its second-moment identity from Weyl orthogonality itself.

The Hostens entry is verified directly on the APS primary record:
Erik Hostens, Jeroen Dehaene, Bart De Moor, Physical Review A 71, 042315,
published 11 April 2005, DOI 10.1103/PhysRevA.71.042315.
https://journals.aps.org/pra/abstract/10.1103/PhysRevA.71.042315
Minor optional bibliographic precision: APS's journal title reads
Stabilizer states and Clifford operations for systems of arbitrary dimensions
and modular arithmetic (without the comma before and). The manuscript uses the
arXiv title with that comma and a journal DOI. This is harmless punctuation;
matching the journal title exactly would be cleaner. No other fix is indicated.

QI EQUATION (27) EQUIVALENCE
Primary source: https://arxiv.org/html/2607.15559v1
Equation (27) has the form
  sum_Y p(x+Y_1) p((YR)_1) p((YR)_2),
where R=[[a,b],[b,-a]], a^2+b^2=-1. Put Z=YR, so R^{-1}=-R and
Y_1=-aZ_1-bZ_2. Then the expression becomes
  sum_(Z_1,Z_2) p(x-aZ_1-bZ_2) p(Z_1) p(Z_2).
For pure input, A=p, and this is the convolution A*A_a*A_b used in the
three-weight branch. Its signs/scalings agree with the manuscript's definition
A_t(x)=A(t^{-1}x). In the p=1 mod4 branch take b=0 and a^2=-1: the unused
variable sums to one and the formula reduces to A*A_a. This involves omitting
a zero-weight register, never evaluating A_0 or inverting zero.

Accordingly, the present wording 'agrees ... after its skew change of variables'
is supported provided the corresponding parameter choices are understood.
Optional clarification for reviewers: explicitly add the change Z=YR and
R^{-1}=-R, and say the two-register case takes b=0. This would make the exact
comparison more transparent, without changing the theorem. The sampler
comparison must not be interpreted as a claim that every pure-learning protocol
needs six copies. The manuscript expressly credits Qi's four-copy primitives,
matching Qi's stated pure-input algorithm (abstract/Theorem 13). No correction
is needed to that caveat.

MIXED SCOPE AND PREDECESSORS
Qi Theorem 7 is pure-input optimizer escape with (p-1)F^6/(4p^3); Section VI
explicitly leaves mixed-input anticoncentration as an extension challenge.
The manuscript's mixed theorem is broader in input scope and its displayed
constant is three times that pure bound in the p=3 mod4 branch. This review
checks the comparison, not whether the stronger bound is proved.
Chen Corollary 6.3 already gives the arbitrary mixed-qubit proper learner;
the manuscript does not claim invention of mixed stabilizer learning.
Allcock supplies the skewed/arithmetic sampling antecedent, including the
square-root-of-minus-one split; the manuscript correctly declines primitive
novelty. Noller supplies related tensor-Weyl measurement precedent, not the
complete optimization learner. Hinsche explicitly anticipates a qudit agnostic
extension using Fourier sampling, but its hidden-group promise is a different
task. The introduction's acknowledgment is appropriate.

The local mixed inequality is presented as the analytic increment; the complete
learner is presented as its consequence using known methods. MUB/design
geometry, symplectic tableau operations, and register arithmetic are not claimed
as original. An optional direct citation to Klappenecker/Roetteler
quant-ph/0502031v2 (Theorem 3) would reinforce the design-identity attribution,
but is unnecessary for correctness when the identity is explicitly derived.

LIMITS
No new exhaustive priority search was undertaken for this manuscript check.
Use literature/FINAL_LITERATURE_REFRESH.txt for the preceding actual Google and
primary-version refresh and its citation-coverage limits. This report does not
certify originality against all literature, the proof, implementation, package
hashes beyond the source above, publication readiness, or external human review.
No material required attribution fix was identified; the two optional precision
improvements above may be applied by the parent authoring agent.
