Accuracy, Interpretability, and Differential Privacy via Explainable Boosting
Harsha Nori, Rich Caruana, Zhiqi Bu, Judy Hanwen Shen, Janardhan Kulkarni
Abstract
We show that adding differential privacy to Explainable Boosting Machines (EBMs), a recent method for training interpretable ML models, yields state-of-the-art accuracy while protecting privacy. Our experiments on multiple classification and regression datasets show that DP-EBM models suffer surprisingly little accuracy loss even with strong differential privacy guarantees. In addition to high accuracy, two other benefits of applying DP to EBMs are: a) trained models provide exact global and local interpretability, which is often important in settings where differential privacy is needed; and b) the models can be edited after training without loss of privacy to correct errors which DP noise may have introduced.
BibTeX
@InProceedings{pmlr-v139-nori21a,
title = {Accuracy, Interpretability, and Differential Privacy via Explainable Boosting},
author = {Nori, Harsha and Caruana, Rich and Bu, Zhiqi and Shen, Judy Hanwen and Kulkarni, Janardhan},
booktitle = {Proceedings of the 38th International Conference on Machine Learning},
pages = {8227--8237},
year = {2021},
editor = {Meila, Marina and Zhang, Tong},
volume = {139},
series = {Proceedings of Machine Learning Research},
month = {18--24 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v139/nori21a/nori21a.pdf},
url = {https://proceedings.mlr.press/v139/nori21a.html},
abstract = {We show that adding differential privacy to Explainable Boosting Machines (EBMs), a recent method for training interpretable ML models, yields state-of-the-art accuracy while protecting privacy. Our experiments on multiple classification and regression datasets show that DP-EBM models suffer surprisingly little accuracy loss even with strong differential privacy guarantees. In addition to high accuracy, two other benefits of applying DP to EBMs are: a) trained models provide exact global and local interpretability, which is often important in settings where differential privacy is needed; and b) the models can be edited after training without loss of privacy to correct errors which DP noise may have introduced.}
}