TAMI-MPC:Trusted Acceleration of Minimal-Interaction MPC for Efficient Nonlinear Inference

TL;DR

This paper introduces TAMI-MPC, a trusted acceleration framework for nonlinear secure multi-party computation that significantly reduces communication and computation costs, enabling real-time privacy-preserving machine learning inference on resource-constrained devices.

Contribution

It redesigns core primitives to minimize interaction, eliminates reliance on oblivious transfer using TEEs, and develops a specialized accelerator for efficient nonlinear inference.

Findings

Achieves up to 4.86x speedup on ResNet-50 inference.

Achieves up to 7.44x speedup on BERT-base inference.

Reduces communication rounds from log(n) to 1 per operation.

Abstract

Secure multi-party computation (MPC) offers a practical foundation for privacy-preserving machine learning at the edge. However, current MPC systems rely heavily on communication and computation-intensive primitives-such as secure comparison for nonlinear inference, which are often impractical on resource-constrained platforms. To enable real-time inference under a resource-constrained platform, we introduce a Trusted Acceleration of Minimal-Interaction MPC framework, TAMI-MPC, for nonlinear evaluation. Specifically, we reduce communication cost by redesigning the core primitives, leaf comparison, and tree merge, reducing the interactive round from log(n) to just 1 per operation. Furthermore, unlike prior work that heavily relies on oblivious transfer (OT), a well-known computational bottleneck, we leverage synchronized seeds inside the TEE to eliminate OT for the vast majority of our designs, along with a correlated-randomness reuse technique that keeps new designs computationally lightweight. To fully realize the potential, we design a specialized accelerator that restructures the dataflow across stages to enable continuous, fine-grained streaming and high parallelism, reducing memory overhead. Our design achieves up to 4.86x speedup on ResNet-50 inference, compared with state-of-the-art CNN frameworks, and achieves up to 7.44x speedup on BERT-base inference, compared with state-of-the-art LLM frameworks.