Shahroz Tariq

In the intricate landscape of Security Operation Centres (SOC), the challenges of alert validation and the persistent issue of alert fatigue have long cast a shadow over operational efficiency. The rapid influx of alerts, coupled with the intricacies of discerning genuine threats from false alarms, poses a formidable challenge to SOC experts. A human-AI collaboration is a promising solution for SOCs. However, to know which task should be automated by AI and which should be delegated for a human-ai collaboration solution is a challenge in itself. In this pitch, Shahroz Tariq delves into the heart of this issue and proposes a rejection learning–based solution called learning to defer.