I recently graduated with my PhD in Computer Science (machine and deep learning focus) from UC Berkeley, advised by Michael I. Jordan. I am interested in optimizing learning systems for complex, multi-objective settings. My research focuses on collaborative learning (e.g. federated learning) and data utility maximization (economics of data sharing, optimal data selection). I am on the industry job market and am excited about applying methodology from optimization and statistics to create useful and safe AI technology.
Prior to my PhD, I studied applied math at Harvard University, after which I took a detour to study classical cello and piano performance at The Juilliard School.
Education
University of California, Berkeley, Ph.D Computer Science, 2019-2024
The Juilliard School, Double M.A. Cello and Collaborative Piano, 2016-2019
Harvard University, B.A. Applied Mathematics, 2012-2016
Publications
Provably Personalized and Robust Federated Learning
Transactions on Machine Learning Research, 2023
Mariel Werner, Lie He, Michael I. Jordan, Martin Jaggi, Sai Praneeth Karimireddy
Defection-Free Collaboration between Competitors in a Learning System
in review, 2024
Mariel Werner, Sai Praneeth Karimireddy, Michael I. Jordan
Privacy Can Arise Endogenously in an Economic System with Learning Agents
Foundations of Responsible Computing, 2024
Nivasini Ananthakrishnan, Tiffany Ding, Mariel Werner, Sai Praneeth Karimireddy, Michael I. Jordan
Collaborative Learning: Aligning Goals and Outcomes
Dissertation for PhD in Computer Science, UC Berkeley, 2024
Mariel Werner
Optimal Data Selection: An Online Distributed View
2022, arXiv
Mariel Werner, Anastasios Angelopoulos, Stephen Bates, Michael I. Jordan
Talks and Presentations
Towards Provably Personalized Federated Learning, Federated Learning Workshop, NeurIPS 2022
Provably Personalized and Robust Federated Learning, AI/ML Seminar Series, Center for Machine Learning and Intelligent Systems, University of California Irvine, November 2023
Provably Personalized and Robust Federated Learning, Federated Learning One World (FLOW) Seminar Series, March 2024
mariel_werner@berkeley.edu