SUPR
Secure Federated Machine Learning
Dnr:

NAISS 2024/22-1204

Type:

NAISS Small Compute

Principal Investigator:

Usama Zafar

Affiliation:

Uppsala universitet

Start Date:

2024-10-01

End Date:

2025-10-01

Primary Classification:

10201: Computer Sciences

Webpage:

Allocation

Abstract

Artificial intelligence (AI) is at the core of modern-day applications. In principle, AI assisted solutions are based on three fundamental building blocks, access to the data which is used to provide use case specific information, a machine learning model, and a training process. Over the years, the focus to make AI a viable solution has shifted among these building blocks. In the beginning, the focus was on mathematical modeling; With the advent of massive datasets, the last two decades were dedicated to improve the training processes; And recently the focus has shifted towards security, privacy and trust based AI assisted solutions. Multiple efforts have been reported both in academia and industry to address the privacy and security challenges in FedML. However, offering a secure and privacy-preserved federated training environment at scale requires further research and development. It is mainly due to the gap between conventional security and privacy solutions and the needs of FedML at scale for both cross-device and cross-silo environments. Within the scope of this project, our focus will be on security and privacy-enhancing techniques for federated machine learning.