Preventing Adversarial Use of Datasets through Fair Core-Set Construction

2019/10/23 by Benjamin Spector, Ravi Kumar, Spector, Benjamin +3
Computer Science · Mathematics · #Adversarial Robustness in Machine Learning #Adversarial system #Artificial Intelligence (cs.AI) #Artificial intelligence #Computer science #Core (optical fiber) #Data mining #Data publishing #Data set #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine learning #Privacy-Preserving Technologies in Data #Programming language #Publishing #Set (abstract data type) #Telecommunications #Theoretical computer science #cs.AI #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1910.10871

published in arXiv (Cornell University) (Cornell University) · 6 pages, 2 figures, NeurIPS 2019 Privacy In Machine Learning Workshop (PriML 2019)

arxiv created 2019/10/24 · openalex publication_date 2019/10/24 · arxiv updated 2019/10/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08

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.

Citations

• Differentially Private Federated Learning: A Client Level Perspective
• Towards Federated Learning at Scale: System Design

Related

• Privacy Partitioning: Protecting User Data During the Deep Learning Inference Phase
• Balancing Privacy, Robustness, and Efficiency in Machine Learning
• No Free Lunch in "Privacy for Free: How does Dataset Condensation Help Privacy"
• Privacy for Free: How does Dataset Condensation Help Privacy?
• Security and Privacy Preserving Deep Learning