Privacy-Preserving Plagiarism Checking

Plagiarism is a pressing issue in academia and industry, where individuals often modify existing content without providing due credit to the original creator and unethically claiming the work to be their own. Existing plagiarism detection tools require code owners to share their code, raising privacy and intellectual property concerns. This paper presents a privacy-preserving protocol for plagiarism detection that eliminates the need for code disclosure during similarity computation. Our protocol builds upon the winnowing-based method for efficient and accurate structural similarity computation. It allows plagiarism detection while preserving code privacy, making it valuable for scenarios where code owners are reluctant to reveal their source code.