BELL INC ( CANADA )

24

generated , and the alarm created a ticket and then people started investigating . But in the case of anomaly detection , you ’ re not relying on those alarms . You notice something unusual in the traffic pattern and you might not know exactly what ’ s wrong but you know something isn ’ t right and you can investigate . And then , of course , you can enhance your algorithms . The paradigm is shifting from detecting issues to looking at the bigger picture and understanding why anomalies are happening .”

Another application for machine learning in telecoms is identifying robocalling and spam calls . “ We ’ ve deployed an algorithm that detects

FEBRUARY 2021