Preventing Adversarial Use of Datasets through Fair Core-Set Construction

Abstract

We propose improving the privacy properties of a dataset by publishing only a strategically chosen "core-set" of the data containing a subset of the instances. The core-set allows strong performance on primary tasks, but forces poor performance on unwanted tasks. We give methods for both linear models and neural networks and demonstrate their efficacy on data.

6 pages, 2 figures, NeurIPS 2019 Privacy In Machine Learning Workshop (PriML 2019)

References in corpus (1)

• Towards Federated Learning at Scale: System Design