CrypTFlow2: Practical 2-Party Secure Inference

Abstract

We present CrypTFlow2, a cryptographic framework for secure inference over realistic Deep Neural Networks (DNNs) using secure 2-party computation. CrypTFlow2 protocols are both correct -- i.e., their outputs are bitwise equivalent to the cleartext execution -- and efficient -- they outperform the state-of-the-art protocols in both latency and scale. At the core of CrypTFlow2, we have new 2PC protocols for secure comparison and division, designed carefully to balance round and communication complexity for secure inference tasks. Using CrypTFlow2, we present the first secure inference over ImageNet-scale DNNs like ResNet50 and DenseNet121. These DNNs are at least an order of magnitude larger than those considered in the prior work of 2-party DNN inference. Even on the benchmarks considered by prior work, CrypTFlow2 requires an order of magnitude less communication and 20x-30x less time than the state-of-the-art.

arXiv commentTo appear at ACM CCS 2020. Code available at: https://github.com/mpc-msri/EzPC

References in corpus (4)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Binarized Neural Networks
• XONN: XNOR-based Oblivious Deep Neural Network Inference
• Secure Computation for Machine Learning With SPDZ

Cited by in corpus (16)

• HyPHEN: A Hybrid Packing Method and Optimizations for Homomorphic Encryption-Based Neural Networks
• NeuJeans: Private Neural Network Inference with Joint Optimization of Convolution and FHE Bootstrapping
• Fusion: Efficient and Secure Inference Resilient to Malicious Servers
• HE-MAN -- Homomorphically Encrypted MAchine learning with oNnx models
• Online Efficient Secure Logistic Regression based on Function Secret Sharing
• PUMA: Secure Inference of LLaMA-7B in Five Minutes
• Unsupervised Information Obfuscation for Split Inference of Neural Networks
• Enhancing Security and Privacy in Federated Learning using Low-Dimensional Update Representation and Proximity-Based Defense
• Sisyphus: A Cautionary Tale of Using Low-Degree Polynomial Activations in Privacy-Preserving Deep Learning
• Hyena: Optimizing Homomorphically Encrypted Convolution for Private CNN Inference
• Masked LARk: Masked Learning, Aggregation and Reporting worKflow
• Accelerating Private Large Transformers Inference through Fine-grained Collaborative Computation
• GALA: Greedy ComputAtion for Linear Algebra in Privacy-Preserved Neural Networks
• PrivQuant: Communication-Efficient Private Inference with Quantized Network/Protocol Co-Optimization
• Privacy-Preserving Screening for Record Linkage
• Round and Communication Balanced Protocols for Oblivious Evaluation of Finite State Machines