Topic: “Challenges in the Verification of Machine Learning Algorithms”
Speaker: Perry van Wesel, NIA Visitor
Date: Friday, December 2, 2016
Location: NASA, Building 1220, Room 27
Time: 1:00pm
Host: A. Goodloe, NASA/LaRC
Abstract: Machine learning (ML) is increasingly being applied to a wide array of domains from search engines to autonomous vehicles. These algorithms, however, are notoriously complex and hard to verify. This work looks at the assumptions underlying machine learning algorithms as well as some the challenges in trying to vary ML algorithms. Furthermore, we focus on the specific challenges of verifying reinforcement learning algorithms. These are highlighted using a specific example. Ultimately, we do not offer a solution to the complex problem of ML verification, but point out possible approaches for verification and interesting research opportunities.
Bio: Perry is a computer science Master student at the University of Eindhoven. During his Bachelor’s degree he had a light focus on artificial intelligence courses, and during his Master’s he specialized in the direction of formal methods. He has been here for nine weeks combining these two topics into his research.