IAS/ Princeton Women and Math Program

IAS/ Princeton Women and Math Program

-
IAS/Princeton Women and Math Program
Fine Hall 314

Register for this event at:
https://math.princeton.edu/princetonday-2022

 

10:00 AM - 11:00 AM
Speaker: Fan Wei, Instructor of Mathematics, Princeton University
"Beyond worst case analysis"

What are some meaningful ways to make sense of how fast an algorithm works? In this talk, we address this question by discussing some theoretical frameworks including smoothed analysis and random graph model, with applications in game theory, graph theory, and social networks. 

11:00 AM - 12:00 PM
Speaker: Barbara Engelhardt, Professor of Computer Science, Princeton University                                                             "Gaussian processes for spatial genomics data analyses"

2:30 PM - 3:30 PM
Speaker: Elizaveta Rebrova, Assistant Professor of Operations Research and Financial Engineering
, Princeton University
"Randomized iterative methods for solving large corrupted linear systems"

There are very few problems that can match the least squares fitting problem in terms of ubiquity of applications. When the data is large, or comes in a streaming way, randomized iterative methods, such as Randomized Kaczmarz method, provide a truly efficient way to search for the least squares solution, or just solve a huge linear system. However, if the labels are corrupted with arbitrarily large sparse errors, the iterates can be diverted far from the solution by each corrupted equation they encounter. 

In this talk, I will discuss our recently proposed robust versions of Randomized Kaczmarz and Stochastic Gradient Descent methods that manage to avoid harmful corrupted equations by exploring the space as they proceed with the iterates. I will also discuss how to use the information obtained on the exploration phase more efficiently, and what structural characteristics of the data matrix are crucial for such methods. 

4:30 PM - 5:30 PM Panel Discussion                                                                                              
"Social and Ethical Implications of Machine Learning"

  • Danqi Chen, Assistant Professor of Computer Science, Princeton University
  • Margaret Holen, Lecturer in Operations Research and Financial Engineering, Princeton University
  • Cynthia Rudin, Professor of Computer Science, Duke University
  • Olga Russakovsky, Assistant Professor of Computer Science, Princeton University