Timezone: »

 
Federating for Learning Group Fair Models
Afroditi Papadaki · Natalia Martinez · Martin Bertran · Guillermo Sapiro · Miguel Rodrigues

@
in New Frontiers in Federated Learning: Privacy, Fairness, Robustness, Personalization and Data Ownership »

Federated learning is an increasingly popular paradigm that enables a large number of entities to collaboratively learn better models. In this work, we study minmax group fairness in paradigms where different participating entities may only have access to a subset of the population groups during the training phase. We formally analyze how this fairness objective differs from existing federated learning fairness criteria that impose similar performance across participants instead of demographic groups. We provide an optimization algorithm -- FedMinMax -- for solving the proposed problem that provably enjoys the performance guarantees of centralized learning algorithms. We experimentally compare the proposed approach against other methods in terms of group fairness in various federated learning setups.

Author Information

Afroditi Papadaki (UCL)
Natalia Martinez (Duke University)
Martin Bertran (Duke University)

I am a PhD student at Duke University. My main research interests are robustness, generalization, and representation learning. My work has focused on robustness in supervised learning in the context of fairness and Pareto efficiency, and on studying the characteristics of good representations for generalization in the context of reinforcement learning.

Guillermo Sapiro (Duke University)
Miguel Rodrigues (UCL)

More from the Same Authors