Protecting the ML Pipeline: Practical Guidance for Securing Machine Learning Systems

Advancements in Machine Learning (ML) have enabled a surge in adoption of ML solutions to address problems across numerous domains. With this rising reliance on ML in many organizations, it is critical that such systems are protected from malicious activities. This talk will present ML-specific cybersecurity issues, discuss ML adversarial techniques, and explore case studies of real-world ML cyber incidents. Further, this presentation will describe secure machine learning systems development approaches and secure machine learning operations (MLOps) pipelines.