Under the hood
Explore how Neraium blends multivariate statistics with deterministic governance to produce reliable infrastructure intelligence.
// Compute multivariate deviation
let diff = sensorVector - meanVector;
let invCov = invert(covarianceMatrix);
let distance = Math.sqrt(diffT * invCov * diff);
// Evaluate against drift threshold
if (distance > threshold) {
if (passesPersistence &&
passesCorrelation &&
withinPhysicalLimits) {
admit();
} else {
suppress();
}
} else {
voidResult();
}
Modelling infrastructure behaviour through relationships
Neraium focuses on how distributed telemetry moves together, making it possible to recognize structural deviation rather than just isolated threshold events.
Multivariate distance
Measure behavioural deviation relative to an established operating geometry.
Covariance structure
Capture how infrastructure signals interact across conditions, assets, and operating states.
Adaptive baselines
Keep detection relevant as valid operational patterns evolve over time.
Detection becomes intelligence only after constraint
Neraium uses deterministic filters to reduce ambiguous or weak conclusions before information is surfaced operationally.
Physical validity
Signals must remain plausible within infrastructure interpretation boundaries.
Persistence
Deviation must persist long enough to indicate meaningful structural change.
Cross-signal agreement
Relevant system changes require corroboration across related telemetry.