Theory of Sufficient Experiments: Target-Selective Identifiability, Robust Recovery, and Minimum-Cost Observation
The Theory of Sufficient Experiments (TSE) develops a mathematical framework for determining the minimum experimental information required to identify, reconstruct, or certify a specific property of an unknown system without necessarily recovering its complete internal state. The theory formalizes target-selective identifiability, observational equivalence, bounded-noise recovery, and minimum-cost experimental design. It establishes an exact minimax criterion based on the separation of critical hypothesis pairs, derives a cost-weighted experiment selection formulation, and connects linear recovery stability to measurement geometry. A central application concerns four-dimensional spacetime curvature. Under explicitly defined holonomy measurements, the study demonstrates that scalar curvature can be recovered from three suitable null-plane probes, while four non-null spatial probes are sufficient and necessary within the corresponding measurement protocol, without reconstructing the full curvature tensor. The framework is examined through mathematical demonstrations and computational experiments spanning computed tomography, synthetic neural dynamics, quantum entanglement certification, nuclear magnetic resonance, fluorescence spectroscopy, industrial sensor calibration, electrical networks, and distributed systems. An independent computational verifier evaluates 3,844 finite experimental configurations, 180 linear optimization instances, 200 algebraic curvature tensors, and additional mathematical cases. The accompanying research package preserves 69 experimental archives, source code, verification procedures, and reproducibility records. Together, these results establish a unified approach to designing experiments according to the property sought, the information available, measurement uncertainty, and operational cost.
Authors
- Luan Rosa
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-15
- DOI
- https://doi.org/10.5281/zenodo.24259370
- Primary Topic
- Optimal Experimental Design Methods
- Type
- preprint