Agama Technologies has introduced AI Anomaly Detection for video service based on artificial intelligence (AI) and machine learning (ML). AI Anomaly Detection is designed to automatically identify anomalies based on information from every subscriber and provide actionable alerts, visualization of detected anomalies and interactive analytics.
"We are excited to introduce the new AI Anomaly Detection feature," said Johan Gorsjo, director of product management at Agama. "Separating actual anomalies from normal variations in KPIs is an excellent example of how AI and machine learning can be applied to video service assurance in a way that addresses real-world needs."
Agama's AI Anomaly Detection employs automated self-learning to recognize patterns in video delivery networks. Acting on information collected in real time from as many as several million client devices, such as set-top boxes and OTT player applications, the algorithm is designed to predict how each subset of the population, from whole countries down to individual neighborhoods, will behave based on past observations.
AI Anomaly detection is designed to help service providers understand where in the delivery chain anomalies occur, what the current situation is, and what has happened before and after the detected anomalies. Detecting real anomalies and putting them into context is intended to creates situational awareness and enable faster analysis and problem resolution.