Private Weighted Sum Aggregation

As large amounts of data are circulated both from users to a cloud server and between users, privately aggregating the shared data is critical. This article considers the problem of private weighted sum aggregation with secret weights , where an aggregator wants to compute the weighted sum of the lo...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on control of network systems 2022-03, Vol.9 (1), p.219-230
Main Authors: Alexandru, Andreea B., Pappas, George J.
Format: Article
Language:English
Subjects:
Agglomeration
Algorithms
Cloud computing
Collaborative work
Computational modeling
Cryptography
data privacy
Differential privacy
distributed systems
Encryption
Games
Multidimensional data
Privacy
security
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:As large amounts of data are circulated both from users to a cloud server and between users, privately aggregating the shared data is critical. This article considers the problem of private weighted sum aggregation with secret weights , where an aggregator wants to compute the weighted sum of the local data of some agents. Based on the privacy requirements posed on the weights, there are different secure multiparty computation schemes exploiting the knowledge structure. First, we review schemes for when each agent has a local private value and local private weight, and when agents have a local private value, but the aggregator holds the corresponding weights. Our main focus is the more general case, where the weights are known neither by the agents nor by the aggregator-they are generated and kept private by a system operator, and the aggregator has to compute the weighted sum without learning the agents' data or the weights. We give solutions that achieve aggregator obliviousness and design more efficient communication and computation strategies for multidimensional data by batching the data into fewer ciphertexts. Finally, we implement our schemes and discuss the numerical results.
ISSN:2325-5870
2325-5870
2372-2533
DOI:10.1109/TCNS.2021.3094788