Natalie Frank (natalief@uw.edu) https://natalie-frank.github.io/ joined the University of Washington as a Pearson and IFDS Fellow in September 2024 working with Bamdad Hosseini and Maryam Fazel. Prior to joining the University of Washington, she obtained her Ph.D. in Mathematics at NYU advised by Jonathan Niles-Weed.
Natalie researches the theory of adversarial learning using tools from analysis, probability, optimization, and PDEs. During her PhD, she focused on studying adversarial learning without model assumptions. Some topics she would like to explore at UW include leveraging convex optimization to enhance adversarial training, using these insights to perform distributionally robust learning for PDEs, and understanding how these topics interact with flatness of minimizers.