Abstract:Today, we are in the era of big data, and data are becoming more and more important, especially private data. Secure Multi-party Computation (SMPC) technology enables parties to perform computing tasks without revealing original data. However, the underlying implementation of SMPC is too heavy, such as garbled circuit (GC) and oblivious transfer(OT). Every time a piece of data is added, the resources consumed by GC and OT will increase a lot. Therefore, it is unacceptable to process large-scale data in a single SMPC task. In this work, we propose a novel theory called SMPC Task Decomposition (SMPCTD), which can securely decompose a single SMPC task into multiple SMPC sub-tasks and multiple local tasks without leaking the original data. After decomposition, the computing time, memory and communication consumption drop sharply. We then decompose three machine learning (ML) SMPC tasks using our theory and implement them based on a hybrid protocol framework called ABY. Furthermore, we use incremental computation technique to expand the amount of data involved in these three SMPC tasks. The experimental results show that after decomposing these three SMPC tasks, the time, memory and communication consumption are not only greatly reduced, but also stabilized within a certain range.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address the issue of excessive resource consumption when handling large-scale data in Secure Multi-Party Computation (SMPC) tasks. Specifically: 1. **Background and Motivation**: - In the era of big data, data is becoming increasingly important, especially private data. - SMPC technology allows parties to perform computational tasks without revealing the original data. - However, traditional SMPC implementations (such as Garbled Circuits GC and Oblivious Transfer OT) are very resource-intensive. With each additional piece of data, the resource consumption of GC and OT increases significantly. - Therefore, handling large-scale data in a single SMPC task is unacceptable. 2. **Proposed New Theory**: - The authors propose a new theory—SMPC Task Decomposition (SMPCTD), which decomposes a single SMPC task into multiple SMPC sub-tasks and multiple local tasks, thereby significantly reducing the consumption of computation time, memory, and communication resources without revealing the original data. - By decomposing operations that satisfy commutativity and associativity in computational tasks, a large SMPC task can be broken down into multiple smaller tasks, thereby improving efficiency and reducing resource consumption. 3. **Specific Applications**: - The authors use this theory to decompose three SMPC tasks in machine learning (Principal Component Analysis PCA, Singular Value Decomposition SVD, and Factor Analysis FA) and implement them through a hybrid protocol framework ABY. - Experimental results show that the decomposed three SMPC tasks not only reduce the consumption of computation time, memory, and communication resources but also stabilize these consumptions within a range. In summary, the goal of this paper is to significantly reduce the resource consumption of SMPC tasks through task decomposition, enabling efficient handling of large-scale data without revealing the original data.

SMPC Task Decomposition: A Theory for Accelerating Secure Multi-party Computation Task

Secure Multi-Party Computation: Theory, practice and applications.

Efficient Secure Computation from SM Series Cryptography

Empowering SMPC: Bridging the Gap Between Scalability, Memory Efficiency and Privacy in Neural Network Inference

A Secure Multiparty Computation Round Optimization Scheme Based on Standard Assumption

Research on Decomposition and Offloading Strategies for Complex Divisible Computing Tasks in Computing Power Networks

Mixed-protocol multi-party computation framework towards complex computation tasks with malicious security

Delay-Optimal Computation Offloading in Large-Scale Multi-Access Edge Computing Using Mean Field Game

Simplifying Multiparty Computation: A Client-Driven Metaprotocol for Federated Secure Computing

Adaptive Privacy-Preserving Coded Computing With Hierarchical Task Partitioning

PPMLAC: High Performance Chipset Architecture for Secure Multi-Party Computation

Secure Multi-Party Computation for Machine Learning: A Survey

SSNet: A Lightweight Multi-Party Computation Scheme for Practical Privacy-Preserving Machine Learning Service in the Cloud

Achieving Secure and Differentially Private Computations in Multiparty Settings

TPD: Temporal and Positional Computation Offloading with Dynamic and Dependent Tasks

A Secure Multi-Party Computation Protocol for Malicious Computation Prevention for preserving privacy during Data Mining

Block-SMPC: A Blockchain-based Secure Multi-party Computation for Privacy-Protected Data Sharing

Efficient maliciously secure two-party mixed-protocol framework for data-driven computation tasks

A Self-Optimizing Computation Partitioning Algorithm for Distributed Many-Task Computing

Task Offloading Optimization in Mobile Edge Computing under Uncertain Processing Cycles and Intermittent Communications

Fully Privacy-Preserving Distributed Optimization Based on Secret Sharing