I worked as a Research Assistant in the Data Security and Privacy Lab, where I worked with Professor Lei Yu on privacy vulnerabilities in machine learning and large language models (LLMs). I also conducted research in Professor Oshani Seneviratne’s BRAIN Lab, focusing on label inference attacks in federated learning.
Projects
- Membership Inference Attacks against ML Models: Investigated potential privacy risks associated with machine learning models, focusing on data leakage and user privacy concerns.
- Membership Inference Attacks against LLMs: A natural extension of the previous project, this research delved into the unique challenges posed by large language models, including their capacity to memorize and regurgitate training data.
- Label Inference Attacks in Federated Learning: Explored vulnerabilities in federated learning frameworks, specifically targeting label inference attacks and their implications for user privacy.
Responsibilities
- Conducted research on privacy vulnerabilities in ML and LLMs, enhancing understanding of data protection and ethical AI practices.
- Designed and implemented experiments to evaluate privacy risks, analyze results, and develop methodologies to enhance privacy vulnerability assessments.
- Performed literature reviews, coded for data analysis and visualization, and contributed to research papers and presentations.