<p>Major: Computer Science<br>
Faculty Mentor: Dr. Yu Xiang Wang, Erchi Wang, Esha Singh<br>
A branch of statistics called "Survival Analysis" makes use of "survival models" that can predict when a certain event will happen, which is especially useful in healthcare to predict critical events like a stroke, heart failure, etc. These models in healthcare, however, make use of sensitive patient data that can easily be exploited and used unethically. This project develops new techniques for training survival models with strong differential privacy guarantees — the gold standard for protecting individual privacy. The goal is to achieve the best privacy-utility tradeoff by leveraging various new techniques developed in modern differentially privacy accounting and deep learning. This will protect and obfuscate the data points of the individuals (like patient data) used in training survival models, while allowing the ability to still derive meaningful results from the data to train the models.</p>