Presented by Nahid Anwar, Computing PhD Cybersecurity emphasis
Location: City Center Plaza 352
The primary objective of this research is to develop an advanced framework for detecting voter registration anomalies, with a specific focus on fraud detection, using the Idaho Voter Registration Election Dataset. The data set contains both anonymized real voter data and synthetically generated fraudulent instances, allowing for a comprehensive examination of potential vulnerabilities in voter registration systems. The real data was obtained from the Idaho Secretary of State’s office. The initial part of the research involved data analysis and identification of misinformation and potential disinformation using statistical analysis and approximate string matching algorithms. Subsequently, we have created the aforementioned anonymized dataset that preserves relationships between attributes in the form of similarity graphs between the major attributes. By leveraging advanced machine learning techniques—such as spectral graph theory, positional embeddings, and deep learning models—the thesis aims to detect and classify fraudulent voter records while preserving the privacy of individuals. Moreover, our research’s approach is designed with flexibility, making it generalizable to voter registration systems in other states. This research introduces the use of similarity graphs for key voter registration attributes, capturing complex interrelationships and patterns within the data. The construction of comprehensive graphs from these attributes, containing over 2 million nodes and 22 million edges, allows for the application of powerful algorithms to identify irregularities that may indicate fraudulent activity. Furthermore, by testing both traditional machine learning models and modern approaches like Graph Neural Networks and Deep Learning, this work seeks to demonstrate the effectiveness of combining these techniques in voter fraud detection. The results of this study will provide election administrators with new tools to identify anomalies and improve the accuracy of voter registration data, thus contributing to the broader goal of strengthening election security and public trust.
Dr. Amit Jain (Chair), Dr. Edoardo Serra, Dr. Francesca Spezzano