Repository navigation
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Result
Every integer width f ≥ 17 admits an exact Gaussian-dyadic left inverse of the normalized full four-qubit Clifford tensor map. This reduces the sufficient width in #7797 from 256 to 17. With the complex threshold of 13 from #7766, the four-qubit dyadic question remains open at widths 13–16. There is no change to κ. Proof · Reproduction and handoff
For each target, a Gaussian-dyadic Bell transform sends the normalized Choi state to a basis vector. The length-16 binary code RM(1,3) ⊕ RM(1,3) cancels the stabilizer phases on each binary subspace. Möbius inversion isolates the target, and a seventeenth coordinate removes affine supports missing the origin. The proof shows that 17 is the exact complex independence threshold for all eight-qubit stabilizer rays. The lower bound uses this larger alphabet. A four-qubit Clifford fixture shows precisely how dropping the extra coordinate breaks this construction.
Using #7772’s spectral bounds, a second exact dyadic inverse has Euclidean operator norm below 3.021. Starting the Richardson iteration from the exact interpolant keeps every iterate a left inverse. At 8,192 iterations the distance to the canonical map is below 2⁻⁷⁸. This allows error control on arbitrary normalized observations while respecting the canonical odd-denominator obstruction of #7792. The physical cost of repeated Gram applications, precision, routing and restoration remains open.
Verification
From the repository root with CPython 3.12 and the standard library:
All six registered runs passed on CPython 3.12.14, including normal and optimized runs of three checkers. The primary checker covers 49 ordered pairs of seven literal four-qubit matrices and all 417,199 subspaces on a full-support fixture. The separate checker covers 9,216 normalized identities for the embedded one-qubit alphabet and expands 1,094 tensor coefficients in two representative rows. The stability checker authenticates the original #7772 receipt, checks the 81 rows and verifies the iteration on a normalized example with odd canonical denominators.
The committed bytes, executable modes, source history, official export and all six runs in a read-only Git-free export pass. The full tensor map and dense stable decoder were not materialized. The spectral identification from #7772 and rank calculation from #7766 are cited inputs. No physical caller, selected multiplier replay or formal-kernel proof is asserted.
Collaboration and attribution
This takes the intermediate-width handoff from #7797. Canonical main was pulled on entry and fetched before submission at
e54523517d0f03e58352e647c403a5af3e525a57. The revision-8 audit card was narrowed to this algebraic decoder question. The selected record and all five framework assumptions keep their existing status. Public claim · Task recordPrepared by apoursh with substantial OpenAI Codex and separate agent assistance. Credit Ali Romman for #7766, #7772 and #7792, Nick Petalas and Ali Romman for the prior phase-witness work in #5902 and #5635, and alephnan and the retained Clifford lineage. The proof credits the classical stabilizer, coding-theory and numerical methods. Apache-2.0.