Lightning Talk: A GNN Based Framework for Kubernetes Security Agents: Threat & Vulnerability...
Lightningat Open AI + Data Forum @ Open Source Summit North America 2022 · on Thursday, 2022-06-23· at 18:05· Lone Star G (Level 3), JW Marriott Austin, 110 E 2nd Street Austin, TX 78701·
summary
A GNN-based Kubernetes security prototype maps live OpenMetrics into a heterogeneous knowledge graph, ingests MITRE and Microsoft threat mappings, and runs online incremental training on streaming cluster data to detect technique-level behaviors. Implementation uses the DGL graph library and produces models, pipelines and simulated-attack datasets for benchmarking. Experiments find heterogeneous link-prediction weak on current tooling, so the project focuses on incremental learning, dynamic resource-aware features and publishable pipelines for reproducible threat detection.
AI-written summary · about
chapters (7)
- 0:00GNN-based Kubernetes security intro
- 1:30Cloud security threats
- 4:11Why GNNs for Kubernetes
- 7:41Threat detection with agents
- 9:51DGL experimentation & training
- 10:53Open-source contributions
- 12:17Summary & future
related talks
31mThreat Modeling for Kubernetes: Enhancing Security Posture in Complex and Regulat...
32mFrom Edge To Insight: Cybersecurity Intelligence With Kubeflow...
37mThreat Modelling: How to Improve Your Kubernetes Security Posture with Threat Mod...
38mThreat Modelling Kubernetes: A Lightspeed Introduction
09mLightning Talk: Sailing Kubernetes Operation with AI Power
13mLightning Talk: AI Meets Kubernetes: Langflow & GROQ in Automated Incident Management
06mLightning Talk: AI-Powered Kubernetes Diagnostics With K8sGPT
06mLightning Talk: Making Sense of Observability with Auto-Discovered Security Policies