Automated Risk Evaluation and Mitigation for AI Systems
summary
neurocat condenses five years of field experience into a software module that automates the systematic risk assessment demanded by the EU AI Act, integrating into the MLOps lifecycle to validate models for automotive, biometrics and critical infrastructure. Claims: measure risk not just average performance, using risk = severity × exposure and targeted test methods that simulate rare hazards via corruptions and adversarial attacks. The platform generates failing scenarios, suggests mitigations and re-tests until thresholds are met.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:30Who is Neurocat
- 1:39What is Neurocat
- 2:35AI Act
- 5:20New Regulations
- 6:52Test Strategy
- 8:28Quality Requirements
- 10:25Performance Measures
- 12:26Risk
- 15:48Risk triage
- 17:00Wrap up
related talks
35mRisks and Mitigations for a Safe and Responsible AI
15mCyber Risk Assessment, IT Asset Management and… AI obviously
24mTechnical Tracks
22mReimagining Risk Management for Modern FinTech with AI
44mAI Meets Cybersecurity: Boosting Speed Without Losing Control
34mReally a New Mouse Trap? Exploring Risks with Artificial Intelligence
36mNot Your Mama's Risk Assessment
08mHuman–AI Collab for Reliable Automotive Control Systems