A privacy-preserving approach for cloud-based protein fold recognition

• Published in: Patterns (New York, N.Y.) Vol. 5; no. 9; p. 101023
• Main Authors: Ünal, Ali Burak, Pfeifer, Nico, Akgün, Mete
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 13.09.2024; Elsevier
• Subjects: cloud-based machine learning; data privacy; machine learning as a service; multi-party computation; privacy preserving machine learning; protein fold recognition; recurrent kernel networks; recurrent kernel networks; privacy preserving machine learning; data privacy; machine learning as a service; cloud-based machine learning; protein fold recognition; multi-party computation
• ISSN: 2666-3899; 2666-3899
• DOI: 10.1016/j.patter.2024.101023

The complexity and cost of training machine learning models have made cloud-based machine learning as a service (MLaaS) attractive for businesses and researchers. MLaaS eliminates the need for in-house expertise by providing pre-built models and infrastructure. However, it raises data privacy and model security concerns, especially in medical fields like protein fold recognition. We propose a secure three-party computation-based MLaaS solution for privacy-preserving protein fold recognition, protecting both sequence and model privacy. Our efficient private building blocks enable complex operations privately, including addition, multiplication, multiplexer with a different methodology, most-significant bit, modulus conversion, and exact exponential operations. We demonstrate our privacy-preserving recurrent kernel network (RKN) solution, showing that it matches the performance of non-private models. Our scalability analysis indicates linear scalability with RKN parameters, making it viable for real-world deployment. This solution holds promise for converting other medical domain machine learning algorithms to privacy-preserving MLaaS using our building blocks. •We propose private machine learning as a service for protein fold recognition•We combine recurrent kernel networks and multi-party computation•Our approach computes the same result as plaintext RKN without compromising privacy•We show its linear scalability to the parameters of recurrent kernel networks In the era of cloud-based machine learning, privacy concerns, especially in medicine, are critical. Protecting the privacy of medical data is essential for maintaining patient trust and complying with regulations. Recognizing protein folds is vital for understanding diseases and developing treatments, but it currently lacks a privacy-preserving solution. We present an approach that secures this process, allowing the use of advanced models without compromising data or model privacy. By maintaining high performance while ensuring privacy, our scalable and efficient solution demonstrates the practicality of secure cloud-based machine learning in healthcare. This work highlights the urgent need for privacy-conscious cloud-based machine learning and aims to inspire further advancements, emphasizing the importance of data privacy in medical applications. This work proposes the first privacy-preserving machine-learning-as-a-service approach for protein fold recognition tasks. It utilizes multi-party computation to perform inference on the query sequence via pre-trained recurrent neural networks. The authors design and implement several efficient multi-party computation building blocks to address the required operations in recurrent kernel networks. They demonstrate its correctness on the Structural Classification of Proteins dataset and the scalability of the solution to various parameters on a synthetic dataset.