Skip to content

Prove the sharp Gaussian-dyadic four-qubit Clifford tensor width is 13 (no κ) - #7799

Open
apoursh wants to merge 1 commit into
CrocSwap:mainfrom
apoursh:codex/clifford-bell-fibres
Open

apoursh wants to merge 1 commit into
CrocSwap:mainfrom
apoursh:codex/clifford-bell-fibres

Conversation

@apoursh

@apoursh apoursh commented Nov 19, 2026

Copy link
Copy Markdown
Contributor

Result

The full normalized four-qubit Clifford tensor map admits an exact left inverse over the Gaussian dyadics at every integer width f ≥ 13. Ali Romman’s #7766 gives rank failure over the complex numbers at width 12. Any lower-width inverse would pad to width 12, so 13 is the minimum. This resolves all four widths left open in #7798. The width-17 minimum for the larger eight-qubit stabilizer alphabet retains its original scope. Proof · Package and reproduction

Six Bell pairs expose the action on Pauli labels. The remaining copy retains a matrix and lets us remove the phase of each Pauli lift. This yields sums over cosets of an explicitly identified eight-element subgroup of Sp(8,2). A dyadic identity isolates each symplectic fibre using all 1,024 characters of the symmetric shear subgroup. A final Pauli trace extracts each original coefficient. The identity continues to hold modulo every odd Gaussian prime. The proof of the upper bound is self-contained.

Using #7772’s spectral enclosure and #7798’s iteration, an exact dyadic decoder at width 13 has Euclidean operator norm below 3.021. This is compatible with the canonical map's odd denominators in #7792. No physical caller or new multiplication exponent κ is claimed.

Verification

CPython 3.12.14, standard library only. All four registered runs passed, normally and under optimized Python.

python3.12 -B scripts/research_package.py --package research/clifford-bell-fibres
python3.12 -B scripts/check_research_commit.py --package research/clifford-bell-fibres --check-source-history --check-modes

The primary checker enumerates all 64 candidate shears at the seed, finds its eight-element stabilizer and verifies the group-algebra identity for every character. The averages reproduce 576 character components directly and 448 with the exact factor 1/4. The separate four-qubit matrix implementation checks the phase-sensitive observation and 6,144 within-fibre coefficient extractions across three coordinate frames. On an independent restricted mixture with 80 nonzero coefficients it recovers all 2,048 coefficients from the actual Bell observations, including 1,968 zeros and deliberate cancellations.

Committed bytes, modes, source history and the official export passed. All four runs passed again from a read-only, Git-free export with an empty PATH. The full tensor matrix, an implemented stable physical decoder, prior spectral suites and the selected multiplier replay were not executed.

Collaboration and credit

Prepared by apoursh with substantial OpenAI Codex and separate agent assistance. Upstream was pulled on entry and fetched again before submission at e54523517d0f03e58352e647c403a5af3e525a57. This is a scoped follow-up to #7798 on the revision-8 audit track. Ali Romman supplied the complex threshold, spectral bounds, Bell-sector analysis and canonical obstruction. Nick Petalas and alephnan contributed the preceding character and Clifford framework. The proof identifies the classical methods and the new recovery construction. Apache-2.0. Collaboration claim · Source pins

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant