Private Adaptive Optimization with Side Information

02/12/2022

∙

Adaptive optimization methods have become the default solvers for many machine learning tasks. Unfortunately, the benefits of adaptivity may degrade when training with differential privacy, as the noise added to ensure privacy reduces the effectiveness of the adaptive preconditioner. To this end, we propose AdaDPS, a general framework that uses non-sensitive side information to precondition the gradients, allowing the effective use of adaptive methods in private settings. We formally show AdaDPS reduces the amount of noise needed to achieve similar privacy guarantees, thereby improving optimization performance. Empirically, we leverage simple and readily available side information to explore the performance of AdaDPS in practice, comparing to strong baselines in both centralized and federated settings. Our results show that AdaDPS improves accuracy by 7.7 privacy-utility trade-offs on large-scale text and image benchmarks.

READ FULL TEXT

Private Adaptive Optimization with Side Information

Differentially Private Adaptive Optimization with Delayed Preconditioners

Private Multi-Task Learning: Formulation and Applications to Federated Learning

Accelerated Federated Learning with Decoupled Adaptive Optimization

Mixed Differential Privacy in Computer Vision

Large Scale Private Learning via Low-rank Reparametrization

Learning under Storage and Privacy Constraints

On Dynamic Noise Influence in Differentially Private Learning

Private Adaptive Optimization with Side Information

Related Research

Differentially Private Adaptive Optimization with Delayed Preconditioners

Private Multi-Task Learning: Formulation and Applications to Federated Learning

Accelerated Federated Learning with Decoupled Adaptive Optimization

Mixed Differential Privacy in Computer Vision

Large Scale Private Learning via Low-rank Reparametrization

Learning under Storage and Privacy Constraints

On Dynamic Noise Influence in Differentially Private Learning