The cloud has changed cyber security forever
summary
A supervised-then-unsupervised ML pipeline named Maggie trains on labeled endpoint and network features, uses feature extraction plus cloud-scale compute and customer feedback to refine labels, and becomes a 100% cloud-based GA service. Early deployment across 33 customers analyzes 1.5 million logs and surfaces 740 infected hosts (≈1.12% false positives), later tuning reduces false positives toward 0.1%. The system anonymizes customer data and produces real-time detection, prevention and cloud-threat intelligence.
AI-written summary · about
chapters (8)
- 0:00Start
- 0:22Machine Learning Evolution
- 1:30Machine Leaming Setup
- 2:40Types of Machine Learning
- 3:31Cybersecurity Tasks and Machine Learning
- 12:12Exposing the stealthiest attacks with 99.9% precision
- 14:06Big data threat intelligence and analysis
- 15:29Using Big Data threat intelligence in Cloud Security
related talks
47mFuture-proofing cloud security: A new operating model
48mCatch me if you can: Building a Web Malware Analyzer using Machine Learning
23mLeave It To The Machines
21mMachine Learning and Cyber Security - Detecting malicious URLs in the haystack
30mData Driven Security
29mAI-Powered Anomaly Detection in Cloud Data Pipelines
11mSecure your cloud in real time
31mBuild threat detection models using ML