The Rise of AI in Anomaly Detection
summary
Claims a practical pipeline that turns raw logs into usable anomaly alerts by combining systematic data collection, feature transforms and an unsupervised AI model that separates normal from abnormal behavior. Uses query-driven preprocessing (DQL queries / key extraction), seasonal baselines to avoid false positives and explicit outlier removal, and applies it to credit card fraud and cybersecurity telemetry. Reports fewer manual triage hours and uncovers previously missed incidents while warning about data poisoning and model refinement needs.
AI-written summary · about
chapters (7)
- 0:00Introduction and background
- 2:12What is anomaly detection
- 4:49Data collection and preprocessing
- 8:58Anomaly detection models
- 13:05Tools and techniques used
- 16:44Real-world applications and benefits
- 18:10Conclusion and Q&A
related talks
29mAI-Powered Anomaly Detection in Cloud Data Pipelines
25mMonitoring AI with AI
29mThresholds Are for Old Threats: Demystifying AI and Machine Learning to Enhance SOC Detection
01h 01mD1T2 - Applying Machine Learning to User Behavior Anomaly Analysis
30mBetter data quality with ML & visualization
43mBeyond Logs: Real-Time Anomaly Detection with Machine Learning
48mHunting Stealth Adversaries with Graphs & AI
34mData Mining for Security Anomalies