Private Multi-Task Learning: Formulation and Applications to Federated Learning
This work addresses privacy concerns in multi-task learning for domains like healthcare and finance, but it is incremental as it builds on existing MTL and differential privacy methods.
The paper tackles the problem of ensuring privacy in multi-task learning for sensitive applications by formalizing client-level privacy via joint differential privacy and proposing an algorithm for mean-regularized MTL, with empirical results showing improved privacy/utility trade-offs on federated learning benchmarks.
Many problems in machine learning rely on multi-task learning (MTL), in which the goal is to solve multiple related machine learning tasks simultaneously. MTL is particularly relevant for privacy-sensitive applications in areas such as healthcare, finance, and IoT computing, where sensitive data from multiple, varied sources are shared for the purpose of learning. In this work, we formalize notions of client-level privacy for MTL via joint differential privacy (JDP), a relaxation of differential privacy for mechanism design and distributed optimization. We then propose an algorithm for mean-regularized MTL, an objective commonly used for applications in personalized federated learning, subject to JDP. We analyze our objective and solver, providing certifiable guarantees on both privacy and utility. Empirically, we find that our method provides improved privacy/utility trade-offs relative to global baselines across common federated learning benchmarks.