Skip to content

Latest commit

 

History

History
100 lines (81 loc) · 6.31 KB

sources.md

File metadata and controls

100 lines (81 loc) · 6.31 KB

Sources and dependencies for deterministic integer factorization

Statement · Complete conditional proof · Verification

Accepted external theorem and exact interface

The selected theorem assumes correctness of the Boolean tensor-contraction result in OpenAI's The Exponential Time Hypothesis is False, dated October 25, 2026. This is an explicit premise of the present conditional result. The theorem statement is in quantitative-reduction.tex, subsection tensor-interface, theorem thm:tensor-contraction. Its construction is in quantitative-contraction.tex.

The retrieved statement was checked on October 28, 2026. It counts Boolean edges, supports succinct uniform tensor-entry algorithms, and allows the literal ring (Z/2^P Z)[V]/(V^ell) with 1<=P,ell<=C*s, including ell=1. The graph has a fixed linear edge budget, polynomially many vertices and maximum degree at most ceil(s^(9/10)). Polynomial local arithmetic fits its entry-computation allowance. The verified reduction supplies every one of these requirements; it invokes no arbitrary coefficient-field extension.

The import metadata records the retrieved theorem text's normalized-content digest. This is a content snapshot, not a pinned Git revision or evidence of theorem correctness. The exact premise is the stated theorem, rather than any future modification of the moving branch URL. No external contraction source code is redistributed.

Linnik's established least-prime result provides deterministic polynomial-bit construction of the exact transform field. A primary source is Xylouris, On the least prime in an arithmetic progression. The progression modulus is polynomially bounded in bit length. Searching its polynomial-sized candidates and checking primality by trial division is therefore polynomial in that parameter; no short-interval prime hypothesis or random prime-selection oracle is required.

Local conditional constructions

Manuscript Role and review scope
Succinct FFT multiplication reduction Initial linear-edge arithmetic verifier and adaptive prefix recovery; full external-interface composition.
One-contraction complete recovery Three-forward-transform root verifier, unique extensions, signed digit-moment separation, saturated gcd refinement and basic literal-ring extraction.
Sidon second moments One contraction at square-root precision and square-root-exponential auxiliary work.
Packed square-zero recovery Characteristic-two algebra, binary regular representation, monotone path bound, guarded literal-ring output, Hankel recovery and deterministic trace splitting.
Scoped parameter obstructions Limitations of the stated precision retuning, full-grid interpolation and recursive-template arguments; no general factorization lower bound.

Independent internal reviewers checked the actual external statement, local network/extraction arguments and exact finite evidence. The source-derived obstruction note additionally states its complete Hermite precision formula and the full-grid/full-packet hypothesis explicitly. Those clarifications do not alter the selected factoring theorem.

General deterministic baselines and significance

Harvey and Hittmeir, A log-log speedup for exponent one-fifth deterministic integer factorisation, primary preprint, proves the general bound

Formula

Their June 2026 revision on large-order elements, Section 1.2, still identifies the positive one-fifth deterministic factoring frontier; its new large-order result is an auxiliary theorem. Gao, Feng, Hu and Pan improve logarithmic factors for balanced semiprimes. That hypothesis cannot be removed when comparing their result with a general-input algorithm.

Under the accepted external premise, the reviewed result replaces a fixed positive numerical-input exponent by a vanishing exponent. This is a substantial classical deterministic asymptotic improvement and motivates its fourth homepage position. This significance assessment is an inference from the stated models and bounds; it is not a literature-wide novelty certification. It is also not a polynomial-time factoring result.

For comparison, the rigorous randomized algorithm of Lenstra and Pomerance, A rigorous time bound for factoring integers has a smaller asymptotic exponent in input bit length. The conditional deterministic reduction does not improve that randomized bound or establish a practical performance claim.

Preserved separate branches and import policy

The English research overview retains the direct one-fifth-exponent candidate, lattice/conic lemmas and scoped direct-template restrictions from the earlier archive. They are not added as prerequisites of the selected conditional proof.

The later archive contributes only its four new manuscripts. There are no fabricated intermediate versions. Every supplied member has its original byte length and SHA-256 identity recorded. Retained members map to canonical text or the integrated English overview; excluded members have no retained text. Formatting and attribution were normalized without silently changing mathematical arguments. The original Chinese summary is integrated into the English overview rather than redistributed as a result.

Attribution: icekylinx and GPT-6 Astra Pro.