AWS re:Invent 2017: GPS: Enhancing Customer Security Using AI/ML on AWS (GPSTEC311)

Опубликовано 28 ноября 2017, 18:57

Data exfiltration—also called data extrusion, data exportation, or data theft—is the unauthorized transfer of data. It is a very serious challenge to business because attackers go after business critical or highly confidential data. Data exfiltration can be done manually by a person, or automated using scripts. Attack sophistication increases by the day. Signature-based techniques to defend against attacks are limited and cannot protect against zero-day attacks. To counter this, we use machine learning (ML) techniques. ML is effective at solving many problems in computer vision, robotics, etc., and is increasingly used in security. Learn an ML technique called anomaly detection, and other state-of-the-art techniques to identify data exfiltration attempts.