DPWeVote: differentially private weighted voting protocol for cloud-based decision-making

Abstract

With the advent of Industry 4.0, cloud computing techniques have been increasingly adopted by industry practitioners to achieve better workflows. One important application is cloud-based decision-making, in which multiple enterprise partners need to arrive an agreed decision. Such cooperative decision-making problem is sometimes formed as a weighted voting game, in which enterprise partners express 'YES/NO' opinions. Nevertheless, existing cryptographic approaches to Cloud-Based Weighted Voting Game have restricted collusion tolerance and heavily rely on trusted servers, which are not always available. In this work, we consider the more realistic scenarios of having semi-honest cloud server/partners and assuming maximal collusion tolerance. To resolve the privacy issues in such scenarios, the DPWeVote protocol is proposed which incorporates Randomized Response technique and consists the following three phases: the Randomized Weights Collection phase, the Randomized Opinions Collection phase, and the Voting Results Release phase. Experiments on synthetic data have demonstrated that the proposed DPWeVote protocol managed to retain an acceptable utility for decision-making while preserving privacy in semihonest environment.

Process modeling for smart factories: using science mapping to understand the strategic themes, main challenges and future trends

“…The authors cited focus on improving performance and agility on the shop floor through a workflow configuration and smart objects integration. Another motor theme with strong centrality is “CLOUD-COMPUTING”, the related studies discuss the application of emerging technologies like cloud technologies in supply chain, marketing processes (Ardito et al , 2019) and business process management (Sill, 2016), seeking to achieve better workflows, performance and decision-making (Fernández-Cerero et al , 2020; Yan et al , 2019). The cluster “BUSINESS-PROCESS-MODELLING-AND-NOTATION” (Figure 6) is also a motor theme with high density and strong centrality.…”

Private rank aggregation under local differential privacy

“…For the LDP protocols in the voting scenario, the LDP-WeVote protocol in [22] is the closest to our solution. However, LDP-WeVote is only designed to the weighted voting game in which the inputs are the binary data and the output is the winner alternative instead of the ranking over all alternatives.…”

Locally private Jaccard similarity estimation

“…Joseph et al introduced a new technique and designed LDP protocol to maintain privacy guarantees over long timescales for the applications of frequency and heavy hitter estimation. For considering a kind of cooperative game theoretic scenario, weighted voting games, Yan et al proposed a LDP protocol DPWeVote to allow the semi‐honest curator to approximately estimate the voting results from the collected perturbed weights and opinions by adopting RR in two separate steps.…”