A Survey and Guideline on Privacy Enhancing Technologies for Collaborative Machine Learning

Abstract

As machine learning and artificial intelligence (ML/AI) are becoming more popular and advanced, there is a wish to turn sensitive data into valuable information via ML/AI techniques revealing only data that is allowed by concerned parties or without revealing any information about the data to third parties. Collaborative ML approaches like federated learning (FL) help tackle these needs and concerns, bringing a way to use sensitive data without disclosing critically sensitive features of that data. In this paper, we provide a detailed analysis of state of the art for collaborative ML approaches from a privacy perspective. A detailed threat model and security and privacy considerations are given for each collaborative method. We deeply analyze Privacy Enhancing Technologies (PETs), covering secure multi-party computation (SMPC), homomorphic encryption (HE), differential privacy (DP), and confidential computing (CC) in the context of collaborative ML. We introduce a guideline on the selection of the privacy preserving technologies for collaborative ML and privacy practitioners. This study constitutes the first survey to provide an indepth focus on collaborative ML requirements and constraints for privacy solutions while also providing guidelines on the selection of PETs.

Research on Data Privacy Protection Strategies Based on Artificial Intelligence

“…This indicates that the future development of this field is closely related to these algorithms, and many scholars are still continuing research in this field. [4,5]…”

2. Architecture Landscape

“…Furthermore, with a cloud-native approach, the RAN and CN architectures can be streamlined, e.g., reduce some complexity by removing multiple processing points for a certain message and removing duplication of functionalities among functions [30]. Cloud-native technologies can enable the creation of cloudlets at the edge of the network, with application-to-application and function-to-function communications, which are capable to satisfy a large number of interconnected assets with flexible mesh topologies.…”

A security-friendly privacy-preserving solution for federated learning

“…Second, the participants (i.e., clients) in the FL setting are able to modify the local model updates to alter the global model maliciously such as performing poisoning attacks [4,5]. A detailed analysis of security vulnerabilities and privacy threats in FL can be found in [6,7]. To overcome the first concern, i.e., sensitive information disclosure, privacy-enhancing technologies (PETs) such as Homomorphic Encryption (HE), Secure Multi-party Computation, Differential Privacy (DP), and Confidential Computing have been proposed [8,9,10], which prevent the server from accessing the original local model updates in cleartext so that the server cannot learn any information about the training data of the clients.…”