Mechanism
The composite admissibility evaluator is disclosed as an architectural primitive. It comprises an observation-ingestion interface that accepts one or more governed observations, an actuation-ingestion interface that accepts a proposed actuation when the evaluator is consumed for actuation admissibility, a multi-factor evidential weight computer that computes an effective evidential weight for each ingested observation, an outcome evaluator that produces one of a plurality of admissibility outcomes, and a lineage-emission interface that emits a governed admissibility-determination observation. Inputs are governed observations rather than separate policy decisions: they may originate from passive environmental markers, active environmental sentinels, cognitive infrastructure agents, operating units, and governance-credentialed contributors.
The multi-factor weight computer integrates a plurality of factors rather than applying a single-factor threshold. The disclosed factors are authority, staleness, modality, dispositional, reputation, integrity, and continuity. Observations from a high-authority contributing device carry high evidential weight and may be treated as substrate conditions, while observations from a low-authority contributing device carry low evidential weight and contribute primarily as advisory input.
The outcome evaluator produces one of six disclosed admissibility outcomes for each ingested observation or proposed actuation. An admit outcome admits the observation into the consuming agent's experiential observation store at the computed evidential weight, or permits the proposed actuation. A gate outcome admits the observation at reduced evidential weight, or permits the actuation subject to additional governance-policy-defined constraints. A defer outcome holds the observation pending corroboration and carries a deferral-expiration parameter, after which the deferral resolves to an admit, gate, or reject outcome. A solicit outcome emits a governed discovery query requesting additional observations to resolve uncertainty. A reject outcome excludes the observation or denies the actuation and carries a rejection-reason classification. An escalate outcome produces a cross-domain escalation upon detection of emergent conditions from the combination of ingested observations.
The lineage-emission interface records the inputs, the evidential weights applied, the factors contributing to each weight, the outcome, and the governance-policy version applied, emitting these as a single governed admissibility-determination observation. A reject outcome's rejection-reason classification is selected from insufficient authority, failed continuity validation, stale observation, failed corroboration, dispositional inconsistency, capability-envelope incompatibility, integrity conflict, or any governance-policy-defined rejection class. Deferred observations held in a deferral queue are subject to continuous re-evaluation as new pertinent observations arrive, and each deferral-queue state change is recorded in the lineage field.
Operating Parameters
The plurality of admissibility outcomes supports governance-policy-configurable response gradations. A deploying authority configures the thresholds, the deferral-expiration parameters, the solicit-emission policy, and the escalation rules independently per deployment domain, per observation type, and per authority level. This governance-policy configurability admits deployment-specific tuning of admissibility behavior without architectural modification, and the disclosure does not fix these parameters to particular numeric values.
A deferred observation awaiting corroboration is promoted to an admit outcome upon receipt of a governance-policy-defined number of corroborating observations within the deferral-expiration window, and is demoted to a reject outcome upon receipt of a governance-policy-defined number of conflicting observations. The number of corroborating or conflicting observations required, and the duration of the deferral-expiration window, are themselves governance-policy-defined rather than fixed.
The evaluator admits new factors, new outcome classes, and new rejection-reason classifications through governance-policy update without architectural modification, rather than embedding fixed evaluation logic. The governance-policy version applied to each determination is recorded in the emitted governed admissibility-determination observation.
Alternative Embodiments
In one embodiment, governed observations are ingested and evaluated by a single co-located evaluator. In a second embodiment, the contributing observations originate from independent contributors, including passive environmental markers, active environmental sentinels, cognitive infrastructure agents, operating units, and governance-credentialed contributors, each carrying its own authority credential, so that no single party need hold authority over all inputs.
A third embodiment exercises the defer and solicit outcomes together. An ingested observation that cannot be admitted on current evidence is held in a deferral queue, and a solicit outcome emits a governed discovery query requesting additional observations pertinent to the ingested observation's region. The deferred observation is then re-evaluated upon receipt of responses, promoting to admit upon a governance-policy-defined number of corroborating observations or demoting to reject upon a governance-policy-defined number of conflicting observations within the deferral-expiration window.
A fourth embodiment exercises the gate outcome, admitting an observation at reduced effective evidential weight or permitting an actuation subject to additional governance-policy-defined constraints, so that an input of partial confidence contributes as advisory rather than substrate.
A fifth embodiment exercises the escalate outcome through the cross-domain observation escalation mechanism. Where governed observations from a plurality of independent sensing domains converge on a common spatial-temporal region and satisfy a governance-policy-defined escalation rule, the evaluator emits a governed escalation observation carrying an escalated classification, the escalated authority credential, and references to each contributing observation, reflecting emergent conditions detectable only from the cross-domain combination.
Composition with Other Properties
Composite admissibility evaluation is the third property of the five-property governance chain. The first property, authority-credentialed observation, supplies inputs each carrying a credentialed source identification evaluated through the authority taxonomy. The second property, evidential weighting in a shared governed observation store, weights observations by authority, sensing-modality reliability, and inter-source consistency, supplying the effective weights the evaluator consumes. The fourth property, governed actuator execution, requires composite admissibility approval before any physical actuation. The fifth property, lineage-recorded provenance, links every observation, evaluation, and action through deterministic lineage.
The five-property chain is recursive: observations generated at any primitive's output feed back into the chain, and actuations emitted from the chain pass through every primitive's governance. An admissibility determination emitted by the evaluator is itself a governed admissibility-determination observation that re-enters the chain. Composite admissibility evaluation also composes with health monitoring: a unit producing observations of failed physical-unclonable-function challenge-response consistency or a triggered tamper-evident seal contributes those as governed observations, which the evaluator can weight toward an integrity-conflict rejection or a reduced-weight gate outcome for determinations referencing that unit.
Prior-Art Distinction
The disclosure distinguishes the composite admissibility evaluator from prior threshold-based admission mechanisms, prior voting-based admission mechanisms, and prior rule-based admission mechanisms in a plurality of respects.
First, the evaluator computes a composite evidential weight integrating a plurality of factors, namely authority, staleness, modality, dispositional, reputation, integrity, and continuity, rather than applying a single-factor threshold. Second, the evaluator produces a plurality of outcomes, namely admit, gate, defer, solicit, reject, and escalate, rather than producing a binary admit-or-reject. Third, the evaluator operates uniformly across all admission contexts within the consuming agent, including admission of governed observations into the experiential observation store, admission of proposed mutations into the planning graph, admission of proposed actuations, and admission of proposed governance-policy updates, rather than providing context-specific admission logic.
Fourth, the evaluator emits a governed admissibility-determination observation recording the complete evaluation provenance, the inputs, the evidential weights applied, the contributing factors, the outcome, and the governance-policy version, rather than producing only a pass-or-fail decision. Fifth, the evaluator admits new factors, new outcome classes, and new rejection-reason classifications through governance-policy update without architectural modification, rather than embedding fixed evaluation logic.
Disclosure Scope
The disclosure encompasses the composite admissibility evaluator as an architectural primitive of the governance chain, including its observation-ingestion and actuation-ingestion interfaces, the multi-factor evidential weight computer and its disclosed factors, the outcome evaluator and its plurality of admissibility outcomes (admit, gate, defer, solicit, reject, escalate), the rejection-reason classification of the reject outcome, the deferral queue and its continuous re-evaluation, and the lineage-emission interface that records the inputs, the evidential weights, the contributing factors, the outcome, and the governance-policy version as a governed admissibility-determination observation.
The disclosure provides that the evaluator operates uniformly across all admissibility determinations within the consuming agent, and that thresholds, deferral-expiration parameters, solicit-emission policy, and escalation rules are governance-policy-configurable per deployment domain, per observation type, and per authority level. New factors, new outcome classes, and new rejection-reason classifications are admitted through governance-policy update without architectural modification, preserving lineage continuity across policy generations.
The disclosure is medium-agnostic, substrate-agnostic, modality-agnostic, and domain-agnostic. A system implementing the composite admissibility primitive over the governance chain and the governed mesh protocol is within the scope of the disclosure regardless of specific sensing modality, signaling medium, computing substrate, or deployment domain, and the disclosure expressly contemplates equivalents and variations consistent with this structural specification.
This article describes subject matter disclosed in U.S. Provisional Application No. 64/049,409.
