The Department of Statistics is proud to announce that three graduate students have been awarded summer Graduate Research Associate (GRA) appointments to pursue advanced research alongside faculty mentors. This initiative, which focuses on both academic discovery and professional development, is made possible by a generous gift from the Lubrizol Corporation.

This summer’s fellows—Jaehoon Kim, Alex Nguyen, and Arushi Vishwakarma—will engage in projects ranging from pharmaceutical safety to the mathematical modeling of social networks.
 

Enhancing Patient Safety Through Better Data Analysis
Jaehoon Kim will collaborate with Dr. Max Russo on the project Evaluating High-Dimensional Regression Approaches for Drug–Drug Interaction Screening. Their work addresses a critical issue in modern medicine: pharmacovigilance. When patients take multiple medications at once, they face potential risks from drug interactions that can cause adverse health events.
Statistically identifying these risks is difficult because the number of possible drug combinations often far exceeds the number of patients in a study. Jaehoon and Max will explore regularization techniques that make estimation and testing possible within high-dimension regression-style models. Over the summer, they will conduct a simulation study to explore which statistical approaches best identify true risks while minimizing "false alarms", as well as compare these methods to simpler approaches that are the current standard practice.

Modernizing Standardized Educational and Psychological Testing Calculations
Alex Nguyen will collaborate with Dr. Sally Paganin on the project Scalable Bayesian Model Selection for Item Response Theory Models. Item Response Theory (IRT) is a framework used to analyze responses from surveys and standardized tests to understand both the difficulty of questions and the traits of the people taking them.
As data sets for these tests grow larger, comparing different statistical models becomes a massive computational challenge that can lead to errors or extreme delays. Alex and Sally are developing a new modular computational strategy that simplifies these complex calculations without losing accuracy. Their approach reduces dimensionality through a hybrid integration that combines analytical and numerical methods. This summer, they will expand on initial simulation results to provide a more robust exploration of the model and computational performance. Their work aims to make educational and psychological measurements more reliable across the board.

Mapping the Dynamics of Social Networks
Arushi Vishwakarma will collaborate with Dr. Dena Asta on the project Spectral Dynamics in Latent Space Models. This research advances the theoretical frontiers of Continuous Latent Space (CLS) network models, which are frameworks for modeling complex networks where the distance between nodes (e.g., people) is determined by their relative positions in a low-dimensional continuous space.
Arushi and Dena aim to explain how latent position distributions influence distributions of the systematic structure of large-scale networks. While much of their work will involve abstract probability and random matrix theory, the project has a wonderful real-world application in anthropology. Arushi and Dena will apply their theoretical findings to a dataset involving cattle herd camps. By studying the dynamics of how these groups choose to live together, the team hopes to better understand the decision-making processes behind social exchange and reciprocity.
 

We are incredibly proud of Jaehoon, Alex, and Arushi. Their work exemplifies the spirit of inquiry and the practical impact of statistical science. We look forward to seeing the fruits of their labor this fall.