I am a research scientist at Google working on federated learning. I received a Ph.D. in applied mathematics from the University of Wisconsin-Madison, and went on to do a postdoc with the wonderful Dimitris Papailiopoulos.
I am interested in practical and efficient optimization methods for machine learning, especially in distributed settings.
I care deeply about open-source software. I actively contribute to and develop multiple libraries for distributed and federated learning experimentation, including TensorFlow Federated, Dataset Grouper, and DrJAX.
In my spare time, I foster dogs and cats, and participate in a Flyball team. I also enjoy baking. You can find recipes I am fond of in my dissertation (no, really).
Publications & Preprints
2024
Fine-Tuning Large Language Models with User-Level Differential Privacy
Z. Charles, A. Ganesh, R. McKenna, B. McMahan, N. Mitchell, K. Pillutla, K. Rush. TF2M@ICML 2024 Workshop.DrJAX: Scalable and Differentiable MapReduce Primitives in JAX
K. Rush, Z. Charles, Z. Garrett., S. Augenstein, N. Mitchell. WANT@ICML 2024 Workshop.Federated Automatic Differentiation
K. Rush, Z. Charles, Z. Garrett. To appear, JMLR.Leveraging Function Space Aggregation for Federated Learning at Scale
N. Dhawan, N. Mitchell, Z. Charles, Z. Garrett, K. Dziugaite. TMLR 2024.
2023
Towards Federated Foundation Models: Scalable Dataset Pipelines for Group-Structured Learning
Z. Charles, N. Mitchell, K. Pillutla, M. Reneer, Z. Garrett. NeurIPS 2023.Convergence of Gradient Descent with Linearly Correlated Noise and Applications to Differentially Private Learning
A. Koloskova, R. McKenna, Z. Charles, K. Rush, B. McMahan. NeurIPS 2023.A Rate-Distortion View on Model Updates
N. Mitchell, J. Ballé, Z. Charles, J. Konečný. ICLR Tiny Papers, 2023.
2022
Federated Select: A Primitive for Communication- and Memory-Efficient Federated Learning
Z. Charles, K. Bonawitz, S. Chiknavaryan, B. McMahan, B. Agüera y Arcas.Motley: Benchmarking Heterogeneity and Personalization in Federated Learning
S. Wu, T. Li, Z. Charles, Y. Xiao, Z. Liu, Z. Xu, V. Smith.Optimizing the Communication-Accuracy Trade-off in Federated Learning with Rate-Distortion Theory
N. Mitchell, J. Ballé, Z. Charles, J. Konečný.Iterated Vector Fields and Conservatism, with Applications to Federated Learning
Z. Charles, K. Rush. ALT 2022.
2021
A Field Guide to Federated Optimization
J. Wang, Z. Charles, Z. Xu, G. Joshi, H. B. McMahan, et al.On Large-Cohort Training for Federated Learning
Z. Charles, Z. Garrett, Z. Huo, S. Shmulyian, V. Smith. NeurIPS 2021.Local Adaptivity in Federated Learning: Convergence and Consistency
J. Wang, Z. Xu, Z. Garrett, Z. Charles, L. Liu, G. Joshi.Convergence and Accuracy Trade-Offs in Federated Learning and Meta-Learning
Z. Charles and J. Konečný. AISTATS 2021.Adaptive Federated Optimization
S. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Konečný, S. Kumar, H. B. McMahan. ICLR 2021.
2020
Advances and Open Problems in Federated Learning
P. Kairouz, H. B. McMahan, et al. (including Z. Charles).On the Outsized Importance of Learning Rates in Local Update Methods
Z. Charles and J. Konečný.
2019
Convergence and Margin of Adversarial Training on Separable Data
Z. Charles, S. Rajput, S. Wright, D. Papailiopoulos.DETOX: A Redundancy-based Framework for Faster and More Robust Gradient Aggregation (arXiv)
S. Rajput, H. Wang, Z. Charles, D. Papailiopoulos. NeurIPS 2019.Does Data Augmentation Lead to Positive Margin? (arXiv)
S. Rajput, Z. Feng, Z. Charles, P. Loh, D. Papailiopoulos. ICML, 2019.A Geometric Perspective on the Transferability of Adversarial Directions (arXiv)
Z. Charles, H. Rosenberg, D. Papailiopoulos. AISTATS, 2019.ErasureHead: Distributed Gradient Descent without Delays Using Approximate Gradient Codes
H. Wang, Z. Charles, D. Papailiopoulos.
2018
ATOMO: Communication-efficient Learning via Atomic Sparsification (arXiv)
H. Wang, S. Sievert, Z. Charles, S. Liu, S. Wright, D. Papailiopoulos. NeurIPS, 2018.Stability and Generalization of Learning Algorithms that Converge to Global Optima (arXiv)
Z. Charles and D. Papailiopoulos. ICML, 2018.Approximate Gradient Coding via Sparse Random Graphs (arXiv)
Z. Charles, D. Papailiopoulos, J. Ellenberg.DRACO: Robust Distributed Training via Redundant Gradients (arXiv)
L. Chen, H. Wang, Z. Charles, D. Papailiopoulos. ICML, 2018.Gradient Coding Using the Stochastic Block Model (arXiv)
Z. Charles and D. Papailiopoulos. ISIT, 2018.Subspace Clustering with Missing and Corrupted Data (arXiv)
Z. Charles, A. Jalali, R. Willett. IEEE Data Science Workshop, 2018.Exploiting Algebraic Structure in Global Optimization and the Belgian Chocolate Problem (arXiv)
Z. Charles and N. Boston. Journal of Global Optimization, 2018.Generating Random Factored Ideals in Number Fields (arXiv)
Z. Charles. Mathematics of Computation, 2018.
2017 and earlier
Algebraic and Geometric Structure in Machine Learning and Optimization Algorithms (link)
Z. Charles. Ph.D. Thesis, University of Wisconsin-Madison, Dec 2017.Efficiently Finding All Power Flow Solutions to Tree Networks
A. Zachariah and Z. Charles. Allerton, 2017.Nonpositive Eigenvalues of Hollow, Symmetric, Nonnegative Matrices (arXiv)
Z. Charles, M. Farber, C. R. Johnson, L. Kennedy-Shaffer. SIAM Journal on Matrix Analysis and Applications, 2013.Nonpositive Eigenvalues of the Adjacency Matrix and Lower Bounds for Laplacian Eigenvalues (arXiv)
Z. Charles, M. Farber, C. R. Johnson, L. Kennedy-Shaffer. Discrete Mathematics, 2013.The Relation Between the Diagonal Entries and the Eigenvalues of a Symmetric Matrix, Based upon the Sign Pattern of its Off-Diagonal Entries
Z. Charles, M. Farber, C. R. Johnson, L. Kennedy-Shaffer. Linear Algebra and its Applications, 2013.