The Cybersecurity Goldilock Problem: Finding Just-Right
summary
Claims that algorithmic thresholds outperform gut and static cutoffs by using Gaussian distributions to define low/medium/high and K‑means for tiering; demonstrates Google Sheets formulas for mean/stddev calculations and a Python runbooks pipeline run via GitHub Actions/Jenkins/cron. Provides concrete numbers (example dataset ~20 apps, mean 165, std 27, three apps flagged high; percentiles like 85% and 5% shown) and a working end‑to‑end prototype that auto‑classifies and records results for audit.
AI-written summary · about
chapters (12)
- 0:00Cybersecurity Goldilocks problem
- 2:10Simplistic rules fail in security
- 6:03Demystifying algorithms
- 8:54Gaussian distribution for risk
- 13:30Visualizing risk with curves
- 17:05Demo with Google Sheets and risk scores
- 20:36Automating risk scoring with CI/CD
- 25:31Tiering apps with K-means clustering
- 31:56Weighted scoring for impact
- 36:57Tiering results and automation
- 40:27Key takeaways and future work
- 41:29Q&A
related talks
23mScalable security: how to win friends and not burn...
45mThe Business Value of Security
23mWhat cyber security can teach us about COVID-19 testing
52mBon Appetit: Establishing an Effective Cyber Risk Appetite
51m2024 Edition: Cybersecurity Standards Scorecard
40mBuilding a Global Cyber Rating - How to Objectively Rate Cyber Capabilities
55mCyber Security vs. Statistics
33mBattle Tested Application Security