- This event has passed.
April 17, 2019 1:00 pm - 2:15 pm EDT
Title: Robust and Reliable Machine Learning: Progress and Challenges
Speaker: Dimitris Tsipras, Massachusetts Institute of Technology
Date: April 17, 2019
Time: 1:00pm – 2:15pm
Location: Northeastern University, 655 Interdisciplinary Science & Engineering Complex (ISEC), 805 Columbus Ave, Boston MA, 02120
Host: Lydia Zakynthinou, Computer Science Theory at Northeastern
Recent progress in machine learning has made the deployment of ML systems in the real world an imminent possibility. But are our current machine learning systems really up to the task? In this talk, I will discuss the brittleness and vulnerability of the existing ML toolkit. I will then describe a conceptual framework that aims to deliver models that are more reliable, and robust to adversarial manipulation. Finally, I will outline how this framework constitutes a new learning paradigm, how it differs from the classic perspective, and what new challenges it gives rise to.
About the Speaker
Dimitris Tsipras is a PhD student in Computer Science at MIT advised by Aleksander Madry. His work focuses on developing efficient tools for building robust machine learning systems and understanding optimization aspects of deep learning. He has spent time as a research intern at Google Brain and as research assistant in NTUA, Greece.