Mr. Karan Alang

Enhancing Network Security with Generalized Isolation Forest and Explainable AI

Abstract:

In today’s hyper-connected digital landscape, traditional rule-based systems often fall short against evolving and stealthy cyber threats. This session introduces a scalable, interpretable framework for network anomaly detection using Generalized Isolation Forest (GIF) — an advanced algorithm tailored for high-dimensional, imbalanced network data.

To bridge the gap between detection performance and transparency, we incorporate ExIFFI (Explainable Isolation Forest Feature Importance) — a lightweight interpretability layer that reveals which features contribute most to each anomaly score. This empowers security teams with faster root cause analysis, actionable insights, and improved incident response.

By combining cutting-edge detection with explainable AI, this approach strengthens automated monitoring systems and builds trust in AI-driven security workflows.

Ideal for security engineers, data scientists, and platform architects, this session equips attendees with a practical blueprint for deploying scalable, explainable anomaly detection systems using GIF and ExIFFI in real-world security pipelines.