Accuracy, Interpretability, and Differential Privacy via Explainable Boosting

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.

arXiv commentTo be published in ICML 2021. 12 pages, 6 figures

References in corpus (7)

• Scikit-learn: Machine Learning in Python
• InterpretML: A Unified Framework for Machine Learning Interpretability
• Gaussian Differential Privacy
• Differential Privacy Has Disparate Impact on Model Accuracy
• Diffprivlib: The IBM Differential Privacy Library
• Differentially Private Bayesian Optimization
• Private Approximations of the 2nd-Moment Matrix Using Existing Techniques in Linear Regression