A Blockchain-based Crowdsourcing Framework

  • Maha Jamal Amin Kadadha

Student thesis: Doctoral Thesis

Abstract

Crowdsourcing is a rapidly growing paradigm that allows individuals and companies to utilize the power of the crowd to fulfill certain tasks. Uber, Amazon MTurk, and UpWork are centralized crowdsourcing frameworks deployed as trusted entities that facilitate the interaction amongst users. Despite their popularity, centralized frameworks may lead to users' privacy breaches and biased execution. While distributed frameworks are proposed in research, trust among the members remains a complex challenge to solve in a fully distributed environment. Ethereum Blockchain has emerged as a trusted infrastructure for the deployment of decentralized applications. Integrating crowdsourcing management as part of Ethereum improves its trustworthiness by enabling transparent end-to-end interactions without a centralized platform. However, there are open challenges that require addressing on the architecture and business levels. In this thesis, blockchain-based crowdsourcing frameworks that can be used by both requesters and workers transparently, autonomously, and cost-efficiently. This thesis consists of four main contributions. In the first contribution, a blockchain-based crowdsourcing framework for sensing tasks is proposed. This framework demonstrates the feasibility of deploying a fully functional blockchain-based crowdsourcing framework with approximate performance to a centralized one paving the way for other contributions. In the second contribution, an auction mechanism on the blockchain is proposed for task allocation to motivate workers to declare truthful costs. In the third contribution, task matching mechanisms that consider the preferences of workers and requesters are proposed. They provide a stable and satisfactory allocation of tasks to workers in a blockchain-based crowdsourcing framework. In the fourth contribution, a lightweight, privacy-preserving machine learning (ML) model is proposed to elevate the capabilities of the blockchain-based crowdsourcing frameworks in improving the reliability of allocated workers and motivate their engagement. All the above contributions were implemented through simulation on a Test Ethereum blockchain using real-world dataset. In summary, this thesis entails multiple frameworks and mechanisms with various properties to answer the varying demands in the domain of blockchain-crowdsourcing.
Date of AwardMay 2021
Original languageAmerican English

Keywords

  • Crowdsourcing
  • Decentralized
  • Task allocation
  • Blockchain
  • Smart Contracts.

Cite this

'