Maziar Sanjabi

Maziar is a research scientist on the AI Integrity team. AI integrity strives to make the world a better place by tackling important problems, such as hate(ful) speech, misinformation and many more, through the power of AI. Prior to joining Facebook, Maziar was at Electronic Arts (EA) where he developed AI models for applications in computer graphics and game design. Maziar completed his PhD at the University of Minnesota, working on optimization methods for statistical signal processing and machine learning. He held postdoctoral positions at UCLA and USC where he worked on scalable AI methods. His current research interests broadly include optimization, multi-modal learning, multi-task and meta-learning, adversarial learning, generative modeling, federated learning, and fairness in machine learning.

Maziar's Publications

May 12, 2020

RESEARCH

ML APPLICATIONS

Fair Resource Allocation in Federated Learning

Federated learning involves training statistical models in massive, heterogeneous networks. Naively minimizing an aggregate loss function in such a network may disproportionately advantage or disadvantage some of the devices. In this work, we propose…

Tian Li, Maziar Sanjabi, Ahmad Beirami, Virginia Smith

May 12, 2020

May 12, 2020

RESEARCH

ML APPLICATIONS

Federated Optimization in Heterogenous Networks

Federated Learning is a distributed learning paradigm with two key challenges that differentiate it from traditional distributed optimization: (1) significant variability in terms of the systems characteristics on each device in the network (systems heterogeneity), and…

Tian Li, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, Virginia Smith

May 12, 2020