Length-twenty spectral purity bound
Abstract
Every feasible rank-three length-twenty spectrum has purity at most 7/121.
Theorem 1.1 (Universal spectral inequality).
Proof. Machine-checked in Lean as D5/S3/Quantum/Entanglement/AbsolutePPT/DimensionTwentySpectralPurityBound.spectral_purity_bound (✓ std3). ∎
Source. Repository-derived.
Commentary.
For every antitone nonnegative real spectrum on Fin 20 with mass one and both rank-three boundary matrices positive semidefinite, the square sum is at most 7/121. The baseline-frozen N6 and N7 rational certificate bounds give the endpoints at lengths 18 and 21. N7 exposes its original private certificate without changing its coefficients or proof. The same-delivery new SpectralBoundaryTransfer.spectral_bound_transfer first gives length 19 using a=2, then length 20 using a=1. Expanding this live inference stops at the baseline endpoints, not at the new transfer. Its arbitrary ordered-tuple construction, actual deletion, repeated insertion and normalization provide the structural content of this upper bound. Length 19 is a spectral intermediate. This inequality does not assert literal four-by-five APPT or settle the unrestricted APPT purity conjecture.
References
- Truth anchor:
D5/S3/Quantum/Entanglement/AbsolutePPT/DimensionTwentySpectralPurityBound.spectral_purity_bound - Dependency: D5/S3/Quantum/Entanglement/AbsolutePPT/QutritN6PurityBound
- Dependency: D5/S3/Quantum/Entanglement/AbsolutePPT/QutritN7PurityBound
- Dependency: D5/S3/Quantum/Entanglement/AbsolutePPT/SpectralBoundaryTransfer