Gold Sponsor Talk: Practical Anomaly Detection At Scale via Self-Supervised Learning

Abstract:
To collect and process log records from various vantage points in compute environments has transformed could security into a big data problem. Despite increasing data volume, however, collecting reliable labeled ground truth data at scale remains to be a major obstacle. As a result, self-supervised learning has emerged as a popular approach for various cloud security applications.  In this talk, we present our experience with building large scale anomaly detection models using self-supervised learning. We discuss challenges of building, evaluating and operating these models at scale to make sure they provide meaningful security value.

SLIDES

About the Speaker:
Baris Coskun is a Senior Principal Scientist at Amazon Web Services. Before joining Amazon Baris held research positions at Yahoo! and AT&T. Baris’ interests include information security, data mining and machine learning.