Cross-Domain Intelligence

Domain-Agnostic Segment Format

200+ entity types share a single compressed representation. Compare a vehicle's trajectory to a network link's behavior — the math is the same.

Unified Representation

One Format, All Domains

Every entity — whether it is a vehicle, a financial instrument, a network link, or a neural organoid — compresses into the same segment structure: a centroid vector, temporal bounds, drift magnitude, and metadata. This uniformity is what makes cross-domain queries possible.

200+ Entity Types

20 Domains, One Index

Autonomous vehicles, finance, networking, spatial computing, IoT, gaming, bio-computing, film/VFX, social dynamics, geospatial, neural emulation, NPC AI, and genomics. Each entity type has a purpose-built ingestion adapter, but downstream — in the index, compression engine, and query planner — they are all segments.

Behavioral Math

Cross-Type Similarity

Because segments share the same vector representation, you can compute cosine similarity between a drone's flight pattern and a stock's price trajectory. If the behavioral shapes match, the similarity score reflects it — regardless of the source domain.

Cross-Domain Cascade Detection

When a regime change in one domain triggers behavioral shifts in another, VectorScaleDB detects the ripple effect automatically.

Detection

Behavioral Ripple Effects

The cascade detector monitors regime transitions across all entity types simultaneously. When entity A in domain X changes regime, the system checks whether correlated entities in domains Y and Z follow suit within a configurable lag window.

Prediction

Historical Lag Analysis

By analyzing historical cascade patterns, the engine predicts cross-domain impacts with timing estimates and confidence scores. These predictions are available via /v1/query/cascade-predict.

Confidence

Statistical Scoring

Every cascade prediction includes a confidence score derived from the number of historical precedents, the consistency of the lag window, and the strength of the behavioral correlation. Low-confidence cascades are flagged but not alerted.

Cross-Domain Alerting Pipeline

From detection to notification in milliseconds. Webhook-driven alerts when behavioral cascades emerge.

Webhooks

Real-Time Notifications

Register webhook endpoints that fire when specific cascade patterns are detected. Each notification includes the source entity, affected entities, cascade type, lag duration, confidence score, and the anomaly scores that triggered detection.

Filtering

Configurable Alert Rules

Set minimum confidence thresholds, domain filters, and entity-type constraints for each webhook. An operations team might only want cascades involving infrastructure entities above 80% confidence; a research team might want everything above 50%.

Composition

Variable-Density Domain Analysis

For any behavioral cluster discovered via variable-density clustering, the domain composition report shows which entity types and domains contribute to the cluster. A cluster of 200 segments might be 40% financial, 35% energy, 25% logistics — revealing cross-domain coupling. Distribution-aware segment comparison measures how regime distributions diverge across domains, while structural anomaly detection surfaces topology shifts invisible to distance-based methods.

Pattern Replay Across Entity Types

Find similar behaviors regardless of domain. Take a known pattern from one entity type and search for it across all others.

Search

Cross-Type Pattern Matching

Submit a trajectory or segment as a query vector and search across all entity types simultaneously. The temporal-semantic index does not discriminate by domain — it returns the nearest behavioral matches regardless of source, ranked by semantic similarity.

Discovery

Emergent Correlations

Patterns that were never explicitly correlated emerge from the data. A supply-chain disruption pattern might match a shipping-route deviation and a commodity-price spike — three domains, one behavioral fingerprint.

Anomaly

Multi-Domain Anomaly Scoring

The anomaly scorer at /v1/query/anomaly-score computes deviation from baselines across all entity types in a single call. An entity that is normal within its own domain but anomalous in the cross-domain context gets flagged.

See It in Action

Detect cross-domain cascades and query anomaly scores across entity types.

# Score anomaly level for an entity across all domain baselines
curl -X POST "https://api.vectorscaledb.com/v1/query/anomaly-score" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "entity_id": "supply-chain-node-42",
    "time_range": {
      "start": "2026-03-09T00:00:00Z",
      "end": "2026-03-09T12:00:00Z"
    },
    "cross_domain": true,
    "include_contributing_domains": true
  }'

# Predict cross-domain cascades from observed regime changes
curl -X POST "https://api.vectorscaledb.com/v1/query/cascade-predict" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "regime_changes": [
      { "domain": "energy", "regime": "high_load", "timestamp_secs": 1772020800 }
    ],
    "predict_from": { "domain": "energy", "regime": "critical_load", "timestamp_secs": 1772022000 },
    "max_lag_secs": 1800
  }'

{
  "cascades_detected": 2,
  "predictions": [
    {
      "target_domain": "manufacturing",
      "predicted_regime": "throttled",
      "estimated_lag_secs": 252.0,
      "confidence": 0.87,
      "supporting_observations": 23
    },
    {
      "target_domain": "logistics",
      "predicted_regime": "rerouting",
      "estimated_lag_secs": 468.0,
      "confidence": 0.76,
      "supporting_observations": 11
    }
  ]
}