AI Reliability in Security - A New Data Science Paradigm

Explore the concept of AI reliability in security through this 25-minute conference talk from the RSA Conference. Delve into the importance of predicting model performance over time in real-world applications. Learn about the key components of AI reliability, methods for maximizing and measuring it, and its necessity in the field. Discover a mathematical algorithm for optimization and gain insights from Celeste Fralick, Chief Data Scientist at McAfee. Cover topics such as software reliability, implementing monitors, and practical applications in data science. Gain a comprehensive understanding of this new paradigm in AI and its implications for security professionals and data scientists.

Introduction
Three words to describe AI
AI reliability
Why AI reliability
AI reliability mathematics
Software reliability
Questions about reliability
Implementing monitors
Review
Monitors
Summary