PrivQuant: Communication-Efficient Private Inference with Quantized Network/Protocol Co-Optimization
This work addresses communication inefficiencies in private inference for server-client systems, offering significant performance improvements but is incremental as it builds on existing 2PC and quantization techniques.
The paper tackles the high inference latency in secure two-party computation (2PC) for private deep neural network inference by proposing PrivQuant, a framework that co-optimizes quantized inference protocols and network quantization, resulting in up to 11x communication reduction and 8.7x latency reduction compared to prior methods.
Private deep neural network (DNN) inference based on secure two-party computation (2PC) enables secure privacy protection for both the server and the client. However, existing secure 2PC frameworks suffer from a high inference latency due to enormous communication. As the communication of both linear and non-linear DNN layers reduces with the bit widths of weight and activation, in this paper, we propose PrivQuant, a framework that jointly optimizes the 2PC-based quantized inference protocols and the network quantization algorithm, enabling communication-efficient private inference. PrivQuant proposes DNN architecture-aware optimizations for the 2PC protocols for communication-intensive quantized operators and conducts graph-level operator fusion for communication reduction. Moreover, PrivQuant also develops a communication-aware mixed precision quantization algorithm to improve inference efficiency while maintaining high accuracy. The network/protocol co-optimization enables PrivQuant to outperform prior-art 2PC frameworks. With extensive experiments, we demonstrate PrivQuant reduces communication by $11\times, 2.5\times \mathrm{and}~ 2.8\times$, which results in $8.7\times, 1.8\times ~ \mathrm{and}~ 2.4\times$ latency reduction compared with SiRNN, COINN, and CoPriv, respectively.