Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

LinkedIn Learning

Security Risks in AI and Machine Learning: Categorizing Attacks and Failure Modes

via LinkedIn Learning

Write review

Overview

Google, IBM & Meta Certificates – 40% Off
One Coursera Plus subscription covers most Professional Certificates on Coursera.
Unlock All Certificates
This course explores building robust, resilient ML systems by addressing intentional adversary attacks and unintentional failures.

Syllabus

Introduction
  • Machine learning security concerns
1. AI Foundations
  • How AI systems can fail and how to protect them
  • Why AI security matters
  • Attacks vs. unintentional failure modes
  • ML security frameworks
  • Security goals for ML: CIA
2. Intentional Failure Modes and Attacks
  • Perturbation attacks and malicious input
  • Poisoning attacks
  • Reprogramming
  • Physical domain: 3D adversarial objects
  • Supply chain attacks
  • Model inversion
  • System manipulation
  • Membership inference and model stealing
  • Backdoors and existing exploits
3. Unintentional Failure Modes and Intrinsic Design Flaws
  • Reward hacking
  • Side effects and misalignment
  • Distributional shifts and incomplete testing
  • Overfitting and underfitting
  • Data bias considerations
4. Building Resilient AI
  • Effective techniques for building resilience in AI
  • Threat modeling AI
  • Dataset threat model
  • Adversarial testing and red teaming
  • API access and supporting components
  • Supply chain
Conclusion
  • Next steps in your AI security journey

Taught by

Diana Kelley

Reviews

4.6 rating at LinkedIn Learning based on 32 ratings

Start your review of Security Risks in AI and Machine Learning: Categorizing Attacks and Failure Modes

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.