Resilient Machine Learning

Do we teach machines correctly? Can they resist cyberattacks learning with only clean data and inside the ideal but not realistic environment?

In this presentation at the Glorium Technologies virtual summit, Dan Shiebler, a Head of Machine Learning at Abnormal Security, tells how to make AI able to work inside a real commercial environment and be less vulnerable to any emergency.

Topic – Resilient Machine Learning

• Maintaining good performance in the face of system outages
• Protecting ML systems from adversaries
• Managing offline/online mismatch
• Protecting against feedback loops in an online system
• Handling new customers and data distribution shift

About speaker

As the Head of Machine Learning at Abnormal Security, Dan Shiebler builds cybercrime detection algorithms to keep people and businesses safe. Before joining Abnormal Dan worked at Twitter: first as an ML researcher working on recommendation systems, and then as the engineering manager for the web ads machine learning team. Before Twitter Dan built smartphone sensor algorithms at TrueMotion and Computer Vision systems at the Serre Lab. Dan’s PhD at the University of Oxford focused on the applications of Category Theory to Machine Learning.

You can get the full recording of the virtual summit «Boosting SaaS products with new technologies. Big Data. Data Science. Artificial Intelligence. Machine Learning» here.