How to Use AI Reliability to Identify and Predict Model Decay

What if we could predict how long our models will last in the field?

Is there a mathematical way to estimate mean time to failure for a specific model?

In this episode, Dave Cole is joined by Celeste Fralick, Chief Data Scientist at McAfee, to discuss AI reliability and how it can help predict model decay.

Celeste also explained:

• What AI reliability measures
• Processes to put in place to measure AI reliability
• The difference between DevOps and MLOps at McAfee
• How adversarial machine learning works
• How to build out a more diverse data science team