Visible to the public The Value of Collaboration in Convex Machine Learning with Differential Privacy

TitleThe Value of Collaboration in Convex Machine Learning with Differential Privacy
Publication TypeConference Paper
Year of Publication2020
AuthorsWu, N., Farokhi, F., Smith, D., Kaafar, M. A.
Conference Name2020 IEEE Symposium on Security and Privacy (SP)
Date Publishedmay
KeywordsCollaboration, composability, Computational modeling, convex machine learning, credit card fraud detection, Data models, data privacy, Differential privacy, differentially-private gradient queries, differentially-private gradients, distributed datasets, distributed private data, financial data processing, financial datasets, fitness cost, fraud, gradient methods, Human Behavior, learning (artificial intelligence), loan interest rates, machine learning, multiple data owners, nonoverlapping training datasets, Prediction algorithms, privacy, privacy budget, privacy-aware data owners, privacy-aware learning algorithms, pubcrawl, regression, regression analysis, Resiliency, Scalability, Stochastic gradient algorithm., stochastic gradient descent, Support vector machines, Training
AbstractIn this paper, we apply machine learning to distributed private data owned by multiple data owners, entities with access to non-overlapping training datasets. We use noisy, differentially-private gradients to minimize the fitness cost of the machine learning model using stochastic gradient descent. We quantify the quality of the trained model, using the fitness cost, as a function of privacy budget and size of the distributed datasets to capture the trade-off between privacy and utility in machine learning. This way, we can predict the outcome of collaboration among privacy-aware data owners prior to executing potentially computationally-expensive machine learning algorithms. Particularly, we show that the difference between the fitness of the trained machine learning model using differentially-private gradient queries and the fitness of the trained machine model in the absence of any privacy concerns is inversely proportional to the size of the training datasets squared and the privacy budget squared. We successfully validate the performance prediction with the actual performance of the proposed privacy-aware learning algorithms, applied to: financial datasets for determining interest rates of loans using regression; and detecting credit card frauds using support vector machines.
Citation Keywu_value_2020