TY - JOUR
T1 - Generating Fair Universal Representations Using Adversarial Models
AU - Kairouz, Peter
AU - Liao, Jiachun
AU - Huang, Chong
AU - Vyas, Maunil
AU - Welfert, Monica
AU - Sankar, Lalitha
N1 - Publisher Copyright:
© 2005-2012 IEEE.
PY - 2022
Y1 - 2022
N2 - We present a data-driven framework for learning fair universal representations (FUR) that guarantee statistical fairness for any learning task that may not be known a priori. Our framework leverages recent advances in adversarial learning to allow a data holder to learn representations in which a set of sensitive attributes are decoupled from the rest of the dataset. We formulate this as a constrained minimax game between an encoder and an adversary where the constraint ensures a measure of usefulness (utility) of the representation. The resulting problem is that of censoring, i.e., finding a representation that is least informative about the sensitive attributes given a utility constraint. For appropriately chosen adversarial loss functions, our censoring framework precisely clarifies the optimal adversarial strategy against strong information-theoretic adversaries; it also achieves the fairness measure of demographic parity for the resulting constrained representations. We evaluate the performance of our proposed framework on both synthetic and publicly available datasets. For these datasets, we use two tradeoff measures: censoring vs. representation fidelity and fairness vs. utility for downstream tasks, to amply demonstrate that multiple sensitive features can be effectively censored even as the resulting fair representations ensure accuracy for multiple downstream tasks.
AB - We present a data-driven framework for learning fair universal representations (FUR) that guarantee statistical fairness for any learning task that may not be known a priori. Our framework leverages recent advances in adversarial learning to allow a data holder to learn representations in which a set of sensitive attributes are decoupled from the rest of the dataset. We formulate this as a constrained minimax game between an encoder and an adversary where the constraint ensures a measure of usefulness (utility) of the representation. The resulting problem is that of censoring, i.e., finding a representation that is least informative about the sensitive attributes given a utility constraint. For appropriately chosen adversarial loss functions, our censoring framework precisely clarifies the optimal adversarial strategy against strong information-theoretic adversaries; it also achieves the fairness measure of demographic parity for the resulting constrained representations. We evaluate the performance of our proposed framework on both synthetic and publicly available datasets. For these datasets, we use two tradeoff measures: censoring vs. representation fidelity and fairness vs. utility for downstream tasks, to amply demonstrate that multiple sensitive features can be effectively censored even as the resulting fair representations ensure accuracy for multiple downstream tasks.
KW - Fair universal representations
KW - algorithmic fairness
KW - generative adversarial networks
KW - minimax games
UR - http://www.scopus.com/inward/record.url?scp=85131706070&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85131706070&partnerID=8YFLogxK
U2 - 10.1109/TIFS.2022.3170265
DO - 10.1109/TIFS.2022.3170265
M3 - Article
AN - SCOPUS:85131706070
SN - 1556-6013
VL - 17
SP - 1970
EP - 1985
JO - IEEE Transactions on Information Forensics and Security
JF - IEEE Transactions on Information Forensics and Security
ER -