Repository navigation
Sharpen leading-section bounds for all completed downsets - #4788
Merged
Mister-Meeseeks merged 1 commit intoOct 23, 2026
Merged
Mister-Meeseeks merged 1 commit into
Mister-Meeseeks merged 1 commit into
Conversation
This was referenced Oct 22, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The completed-downset norm bound is a premise of the Full physical-cut and precision arguments. This package independently proves a stronger bound from the original leading-section conditioning, for every complex matrix and every binary tree of contiguous intervals, including unbalanced trees. With
Λ,Q≥1,||T||₂≤Λand every leading inverse bounded byQ, all completed atomic downsets and their inverses are bounded by32 N^(3/2)(ΛQ)^8, improving the retained sufficient256 N³(ΛQ)^12envelope. The proof restricts one global unit-LDU factorization and keeps the noncommuting frontier orderP=L_C D_I U_F^-1 Uexplicit.An independent exact oracle checks 11 interval trees and 398 downsets, including
N=1, asymmetric trees and eight states with both completed lower factors and a nontrivial upper frontier. An algebra-only execution of the retained scheduler binds its ten size-four atoms. Five controls reject wrong factor order, a singular leading section, noncontiguous inverse restrictions, a whole-inverse-only premise and an unmatched complex adjoint. These fixtures support the written quantified proof; they do not substitute for it.The result concerns complete ideal calls. Unfinished FFT/generator fragments, rounded execution, physical costs and substituting the new envelope into every downstream allocation remain separate obligations. No source allocation or selected frontier is changed. This follows #4755's explicit inherited-norm boundary and differs from #4209's near-identity majorant premises.
Validation: metadata/source pins; registered Python 3.12 normal and optimized replay; committed-byte/source-history preflight (10 package/input files); official export integrity; Git-free normal/optimized replay with all 11 exported files unchanged. No Lean, Docker or full historical replay. Proof and limits:
research/leading-section-downset-audit/README.md.Canonical claim: #4202 (comment)
Submission uses canonical base
624de9dc66858af0945e5f8a0942488d62280d43, priority 7. The 19:49 UTC complete refresh covered 299 open PRs, 3721 Discussion records and 6319 base local notes; scoped search and live #4202 plus relevant predecessor heads were rechecked. Search does not cover unpublished work or all historical code diffs and is not a novelty certificate.Credit Jose Manuel's original scalar-LDU/frontier argument and polynomial bound, Vlado's scheduler, and classical block elimination/Schur/triangular norm methods. The contribution is the independent restriction-based sharpening and arbitrary interval-tree extension. Prepared for Brian (
JoyboyBrian) with substantial OpenAI Codex assistance. Agent cross-review is part of the author process, not independent human/formal acceptance. Hosted CI is not yet claimed.