With the rapid development of cloud computing, clients and users with limited computing resources can outsource their computation-intensive tasks to the Cloud Service Providers (CSPs). However, as the CSPs are commercial in nature and aim to increase their profits, some security challenges are still attached to them. In this paper, we propose an efficient publicly verifiable computation scheme (EPVM) for large-scale matrix multiplication with privacy preservation. Based on the theory of discrete logarithm problem and the techniques of privacy-preserving matrix transformation, our scheme not only protects the privacy of the client’s matrices but also significantly reduces the computation overhead on the client end as well as the CSP side. Our detailed security analysis and proofs show that the proposed scheme can achieve the established security requirements. The experimental evaluation also demonstrates that the proposed scheme works efficiently as compared to other existing solutions.

随着云计算的快速发展，计算资源有限的客户和用户可以将计算密集型任务外包给云服务提供商（CSP）。然而，由于CSP本质上是商业性的，并且以增加利润为目的，因此仍然面临一些安全挑战。在本文中，我们提出了一种有效的公共可​​验证计算方案（EPVM），用于具有隐私保护的大规模矩阵乘法。基于离散对数问题理论和隐私保护矩阵变换技术，我们的方案不仅保护了客户端矩阵的隐私，而且显着减少了客户端和CSP侧的计算开销。我们详细的安全分析和证明表明，所提出的方案可以达到既定的安全要求。实验评估还表明，与其他现有解决方案相比，所提出的方案工作效率更高。