Privacy-Preserving Interest-Ability Based Task Allocation in Crowdsourcing.

Jialu Hao,Cheng Huang,Guangyu Chen,Ming Xian,Xuemin (Sherman) Shen

DOI: https://doi.org/10.1109/icc.2019.8761188

2019-01-01

Abstract:Numerous crowdsourcing applications have emerged in our daily lives, which enable customers to outsource their complicated tasks to a crowd of workers. However, the information of task tags and worker profiles is explicitly obtained by the crowdsourcing server to recommend tasks effectively, which violates the privacy of both customers and workers. Moreover, the worker's ability to do the task should also be verified in a privacy-preserving way. To address these issues, we propose a privacy-preserving interest-ability based task allocation scheme in crowdsourcing, which protects both task and worker privacy and enables the crowdsourcing server to allocate tasks in a fine-grained way. Specifically, by utilizing attribute-based encryption (ABE) and proxy re-encryption based searchable encryption (PRE-SE) on the task content and task tags respectively, customers are able to enforce fine-grained ability requirements on their tasks, and workers can specify flexible interests to choose their desired tasks. Additionally, ElGamal signature enables workers to prove their abilities to the crowdsourcing server without revealing the task content. Numerical analysis and experiment results demonstrate that our proposed scheme is efficient in terms of computation and storage overhead and is practical to be implemented in crowdsourcing.