Private Machine Learning in TensorFlow using Secure Computation

Abstract

We present a framework for experimenting with secure multi-party computation directly in TensorFlow. By doing so we benefit from several properties valuable to both researchers and practitioners, including tight integration with ordinary machine learning processes, existing optimizations for distributed computation in TensorFlow, high-level abstractions for expressing complex algorithms and protocols, and an expanded set of familiar tooling. We give an open source implementation of a state-of-the-art protocol and report on concrete benchmarks using typical models from private machine learning.

References in corpus (1)

• Scalable Private Learning with PATE

Cited by in corpus (16)

• Secure Evaluation of Quantized Neural Networks
• SEALion: a Framework for Neural Network Inference on Encrypted Data
• Enabling Homomorphically Encrypted Inference for Large DNN Models
• A Scalable Approach for Privacy-Preserving Collaborative Machine Learning
• Evaluating Privacy-Preserving Machine Learning in Critical Infrastructures: A Case Study on Time-Series Classification
• nGraph-HE: A Graph Compiler for Deep Learning on Homomorphically Encrypted Data
• Secure Quantized Training for Deep Learning
• Hijack Vertical Federated Learning Models As One Party
• CryptoSPN: Privacy-preserving Sum-Product Network Inference
• Secure Medical Image Analysis with CrypTFlow
• Privacy-preserving Machine Learning for Medical Image Classification
• Efficient CNN Building Blocks for Encrypted Data
• Trustworthy AI Inference Systems: An Industry Research View
• Privacy-preserving collaborative machine learning on genomic data using TensorFlow
• Corella: A Private Multi Server Learning Approach based on Correlated Queries
• Towards General-purpose Infrastructure for Protecting Scientific Data Under Study