Graduate Defense: Akshey Sigdel

Thesis Information

Title: THEMIS: A Fair And Privacy-Preserving Outsourcing Platform Using Blockchain

Program: Master of Science in Computer Science

Advisor: Dr. Gaby Dagher, Computer Science

Committee Members: Dr. Eric Henderson, Computer Science and Dr. Nasir Eisty, Civil Engineering

Abstract

Outsourcing systems rely on the strategic selection of reputable, reliable, and trustworthy workers. Achieving fairness in task allocation is essential, as it directly correlates with increased network participation. Ensuring fairness requires evaluating a worker’s reputation based on their past performance and behavior without compromising privacy. However, directly disclosing a worker’s past performance can compromise privacy, making them vulnerable to targeted attacks, undue scrutiny, and potential biases. Furthermore, while transparency in performance assessments is crucial, it may incentivize malicious or self-serving behaviors, jeopardizing the integrity of the system.

In this thesis, we present THEMIS, a blockchain-based outsourcing platform that utilizes a differentially private graph-based algorithm to select a subset of workers while preserving reputation privacy. THEMIS employs a privacy-preserving protocol to account for individual worker contributions without revealing their performance. Our experiments demonstrate the effectiveness of the proposed solution in selecting reputable nodes while preserving overall privacy. The findings confirm that the proposed algorithm achieves fairness, scalability, and efficiency.