Rotating Observation Kernels and Birth-Order Information Recovery in a Reversible Plücker–Markov Experiment
Support birth in a network does not by itself specify which rate directions a noisy experiment can distinguish. We give a quantitative connection between Plücker boundary geometry, reversible resolvent sensitivity and nuisance-adjusted Fisher information. In a 61-state, 435-edge model, twenty Gaussian readouts retain seventeen geometric unknowns and a shared clock, and one additional readout is selected by a recorded, parameter-independent label. All 93 birth edges gain in their leading information coefficients, although the normalized joint quotient rank increases only from two to three. For an explicit three-edge family, the old mean map has a local equal-response curve. The additional readout is transverse to this curve and restores local identifiability for small positive δ. Its effect on information order is explained by a rotating old kernel: one fixed exact contrast has old information of order δ⁶ and new information of order δ⁴, while a direction following the old kernel remains exactly blind to the old experiment. The old joint three-target matrix still has one exact zero eigenvalue; the sixth-order information is directional. Both absolute information levels vanish at the wall. Exact rational certificates show how birth amplitudes, network voltages and shared calibration jointly determine the effective observation orders. The construction provides a benchmark for studying readout choice near singular boundaries with the sharing of nuisance parameters included in the experimental design.
Authors
- Bin Du
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-21
- DOI
- https://doi.org/10.5281/zenodo.24420449
- Primary Topic
- Gene Regulatory Network Analysis
- Type
- preprint