TY - GEN
T1 - Differentially Private Normalizing Flows for Synthetic Tabular Data Generation
AU - Lee, Jaewoo
AU - Kim, Minjung
AU - Jeong, Yonghyun
AU - Ro, Youngmin
N1 - Publisher Copyright:
© 2022, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.
PY - 2022/6/30
Y1 - 2022/6/30
N2 - Normalizing flows have shown to be a promising approach to deep generative modeling due to their ability to exactly evaluate density - other alternatives either implicitly model the density or use approximate surrogate density. In this work, we present a differentially private normalizing flow model for heterogeneous tabular data. Normalizing flows are in general not amenable to differentially private training because they require complex neural networks with larger depth (compared to other generative models) and use specialized architectures for which per-example gradient computation is difficult (or unknown). To reduce the parameter complexity, the proposed model introduces a conditional spline flow which simulates transformations at different stages depending on additional input and is shared among sub-flows. For privacy, we introduce two fine-grained gradient clipping strategies that provide a better signal-to-noise ratio and derive fast gradient clipping methods for layers with custom parameterization. Our empirical evaluations show that the proposed model preserves statistical properties of original dataset better than other baselines.
AB - Normalizing flows have shown to be a promising approach to deep generative modeling due to their ability to exactly evaluate density - other alternatives either implicitly model the density or use approximate surrogate density. In this work, we present a differentially private normalizing flow model for heterogeneous tabular data. Normalizing flows are in general not amenable to differentially private training because they require complex neural networks with larger depth (compared to other generative models) and use specialized architectures for which per-example gradient computation is difficult (or unknown). To reduce the parameter complexity, the proposed model introduces a conditional spline flow which simulates transformations at different stages depending on additional input and is shared among sub-flows. For privacy, we introduce two fine-grained gradient clipping strategies that provide a better signal-to-noise ratio and derive fast gradient clipping methods for layers with custom parameterization. Our empirical evaluations show that the proposed model preserves statistical properties of original dataset better than other baselines.
UR - http://www.scopus.com/inward/record.url?scp=85142151616&partnerID=8YFLogxK
U2 - 10.1609/aaai.v36i7.20697
DO - 10.1609/aaai.v36i7.20697
M3 - Conference contribution
AN - SCOPUS:85142151616
T3 - Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI 2022
SP - 7345
EP - 7353
BT - AAAI-22 Technical Tracks 7
PB - Association for the Advancement of Artificial Intelligence
T2 - 36th AAAI Conference on Artificial Intelligence, AAAI 2022
Y2 - 22 February 2022 through 1 March 2022
ER -