Luminary Research Archive
Luminary Research ArchiveIndependent scholarship
Luminary Portal

All Research

Research

154 papers
Browse papers
  1. Positive Operator Cocycles in Shrinking CorridorsScott SchubrEdition 2 · 2026-10-31 Public
  2. Coherence thresholds for optimal-rate stabilizer purity testingScott SchubrEdition 2 · 2026-10-31 Public
  3. Spectral escape for interacting and coalescing signed potential formsScott SchubrEdition 5 · 2026-10-30 Public
  4. Large negative subspaces for positive Euler deformationsScott SchubrEdition 3 · 2026-10-30 Public
  5. Fourier computation of empirical diffusion metricsScott SchubrEdition 2 · 2026-10-30 Public
  6. Purity frontiers and optimal copy complexity for qutrit stabilizer measurementsScott SchubrEdition 2 · 2026-10-30 Public
  7. Polynomial cores and inertia under Carleman moment boundsScott SchubrEdition 3 · 2026-10-30 Public
  8. Collision-uniform kernel asymptotics for Báez–Duarte zero packetsScott SchubrEdition 4 · 2026-10-30 Public
  9. Mixing-weight entropy concavity near the Gaussian below one-half signalScott SchubrEdition 3 · 2026-10-30 Public
  10. Covariance-Spectrum Certificates for Homogeneous Stable Count LawsScott SchubrEdition 8 · 2026-10-30 Public
  11. Relative tail payment and spectral enclosuresScott SchubrEdition 1 · 2026-10-29 Public
  12. Joint inclusion bounds for random greedy matroid basesScott SchubrEdition 1 · 2026-10-29 Public
  13. Divisor tests and coefficient transfers for sextic character momentsScott SchubrEdition 1 · 2026-10-28 Public
  14. Strong Self-Permanent Inequalities for Complex Four-Row Bipartite SupportsScott SchubrEdition 3 · 2026-10-28 Public
  15. Signed Frequency Certificates and Constraint Costs for Odd Weil FormsScott SchubrEdition 1 · 2026-10-27 Public
  16. Frequency-Source Resolvent Limits and Positive-Receiver InstabilityScott SchubrEdition 1 · 2026-10-27 Public
  17. Source-Resolved Resolvent Growth under Reciprocal-Theta ConfinementScott SchubrEdition 1 · 2026-10-27 Public
  18. Distributional traces and central moment tests for Vámos quarticsScott SchubrEdition 1 · 2026-10-27 Public
  19. Exact-rank quadratic correlation: a phase-compatible Clique reductionScott SchubrEdition 1 · 2026-10-27 Public
  20. Certified resolvent detection of the third Weyl-blind twisting harmonicScott SchubrEdition 2 · 2026-10-27 Public
  21. Degree-eight positivity and polynomial obstructions for monotone Riemann theta averagesScott SchubrEdition 7 · 2026-10-27 Public
  22. Independent Squares and Average-Radius Obstructions for Linear CodesScott SchubrEdition 2 · 2026-10-27 Public
  23. An explicit quintic obstruction for a fixed-center Riemann theta productScott SchubrEdition 2 · 2026-10-27 Public
  24. A rational reconstruction of a quintic obstruction for a fixed-center Riemann theta productScott SchubrEdition 2 · 2026-10-27 Public

Power Concavity of Low-Dimensional Unimodular Fiber Counts

Original PDF · First-page preview

Loading original first page…

Abstract

We prove a corrected fourth-root Jensen inequality for unimodular integer fibers whose central affine dimension is at most three. For an integrable integer-valued demand A with feasible fibers almost surely and integral mean b, a bounded nonempty central fiber satisfies E[(N(A)−1/12)^(1/4)]≤(N(b)-1/12)^(1/4). This implies a 2/11-power inequality and logarithmic Jensen, with finite expectations and countable support. The proof constructs a monotone continuous extension of type-A₃ lattice counts with all twelve directed bounds varying. A quartic Lorentzian certificate gives the corrected fourth root; shortest-path closure and a common integral rank-three reduction transfer it to arbitrary unimodular systems. For this specified continuous count extension, 2/11 is the largest concavity power and 1/12 is the least fourth-root correction. Endpoint-count thresholds give stronger powers tending to 1/4. We also prove sharp corrected square-root Jensen in dimension two and its optimal leading large-count exponent deficit. For the complete acyclic four-vertex digraph, the classical continuous count extension has optimal power 4/13, while exact integral examples obstruct cube-root Jensen. Directed rounding and a full-dimensional local obstruction explain the role of a global continuous extension.

Read the full paper in the original PDF.