Secure Machine Learning in the Cloud Using One Way Scrambling by Deconvolution

TL;DR

This paper introduces OWSD, a deconvolution-based scrambling method that enhances data privacy in cloud machine learning, achieving near-perfect classification with reduced computational costs compared to traditional encryption methods.

Contribution

The paper presents OWSD, a novel scrambling framework that offers a secure and efficient alternative to homomorphic encryption for cloud-based machine learning.

Findings

OWSD achieves near-perfect classification accuracy.

OWSD significantly reduces computational overhead.

The approach demonstrates robustness against data attacks.

Abstract

Cloud-based machine learning services (CMLS) enable organizations to take advantage of advanced models that are pre-trained on large quantities of data. The main shortcoming of using these services, however, is the difficulty of keeping the transmitted data private and secure. Asymmetric encryption requires the data to be decrypted in the cloud, while Homomorphic encryption is often too slow and difficult to implement. We propose One Way Scrambling by Deconvolution (OWSD), a deconvolution-based scrambling framework that offers the advantages of Homomorphic encryption at a fraction of the computational overhead. Extensive evaluation on multiple image datasets demonstrates OWSD's ability to achieve near-perfect classification performance when the output vector of the CMLS is sufficiently large. Additionally, we provide empirical analysis of the robustness of our approach.