Glossary
Anomaly Detection
Identifying artificial patterns in traffic (bots, fake engagement) through vector space anomalies, not keywords.
Definition
Anomaly Detection identifies patterns that deviate from normal behavior-in Sentient OS, primarily artificial patterns like bot traffic, fake engagement, and manipulation. The Integrity Layer applies vector-space anomaly detection: behavioral vectors that cluster abnormally, engagement patterns that don't follow organic dynamics, growth curves that are too steady (e.g., 2.7%/month flags manipulation). Unlike rule-based detection (which adversaries adapt to), anomaly detection finds statistical outliers in high-dimensional space. The Social Reliability Index and Audience Authenticity Score surface inorganic spikes, bought likes, and bot networks. Anomaly detection protects your investment from corrupted metrics and ensures decisions are based on real signals.
Why It Matters
Anomaly detection protects Sentient's outputs from manipulation. The Integrity Layer ensures you're optimizing for real audiences, not bots or bought engagement.
Related Pages
Related Terms
Bot Detection
Identifying automated fake accounts. Part of the Integrity Layer's Audience Authenticity Score.
Vector Spaces
High-dimensional mathematical spaces where actors, products, campaigns become 'Persona Vectors.' Mathematics instead of databases.
Engagement Rate
Metric measuring interaction depth. Sentient analyzes engagement stability to unmask inorganic spikes.
Integrity Layer
Command Center Module III - the shield against fraud. Social Reliability Index and Audience Authenticity Score protect every recommendation.
Machine Learning
AI systems that learn patterns from data. Sentient uses unsupervised learning for archetypal discovery.
Explore the Full Platform
See how these concepts come to life inside Sentient OS.