Audit and error-correction benchmarks for a ternary tensor model
Audit of the duality-projected ternary tensor model and an exact quantum-error-correction assessment. The sampler audit fixes a large-volume boundary overflow, restricts midpoint measurements to compatible full-step grids, and validates the transfer and conditional updates with independent tests. Autocorrelation reanalysis and two longer independent 4³ runs expose remaining uncertainty in curvature-regime weights. A control variate from exact boundary duality gives held-out mean-curvature estimates of 2.9703(26) and 2.9719(26), conditional on stationary sampling. The original constraints define a qutrit CSS code with a weight-four logical Pauli at every lattice side length of at least two. Exact searches give [[48,20,2]]₃, [[162,64,3]]₃, and [[384,132,4]]₃ on sides two, three, and four. An ideal-syndrome lookup corrects all 3,072 single-qutrit Pauli errors at side four. A separate curvature-check encoding gives [[48,6,2]]₃, [[162,18,3]]₃, and [[384,6,4]]₃ in those finite tests and admits a commuting-projector memory Hamiltonian. Scalability, noisy fault tolerance, quantum advantage, and graviton dispersion remain unresolved. The archive contains corrected code, regression tests, raw traces, exact code benchmarks, and the editable manuscript.
Authors
- Sushabhan Giri (ORCID: https://orcid.org/0009-0006-1579-5477)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-17
- DOI
- https://doi.org/10.5281/zenodo.24317914
- Primary Topic
- Quantum Computing Algorithms and Architecture
- Type
- preprint