NSF Awards GRFP to Four Harvard Statistics Students

When this year’s Graduate Research Fellowship Program (GRFP) awardees were announced, we were thrilled to hear that four out of six recipients in the “Mathematical Sciences – Statistics” category were Harvard Department of Statistics students! This outstanding performance is a testament to our students’ ability to craft research proposals that stood out in terms of their intellectual merit and potential for broad impacts on society.

Started in 1951, the GRFP program was implemented to help grow future generations of scientists in the US and has supported over 70,000 fellowships (source: https://www.nsfgrfp.org/about/about-grfp/). According to the National Science Foundation’s website, the “GRFP recognizes and supports outstanding graduate students who have demonstrated the potential to be high achieving scientists and engineers, early in their careers.”

To highlight our students’ achievement in receiving this prestigious fellowship, we reached out to them – 2nd year PhD students Alan Chung, Nathan Cheng, and Matthew Esmaili Mallory as well as senior concentrator Skyler Wu – to learn more about their research proposals. Honored to receive this 3-year fellowship, our students shared more about their research areas of interest (lucky for us!).

Nathan Cheng on Selective and Causal Inference:

““I am interested in statistical methodology that addresses new problems introduced by modern datasets and statistical needs. Currently, I am working on a project in selective inference that explores questions such as, “How can one use the same data to both determine hypotheses of interest and validly test these hypotheses?”  In addition, I am interested in sensitivity analysis in causal inference.  In a sensitivity analysis, one examines how causal conclusions can vary, strengthen, or evaporate as we add and remove assumptions about the data.  I am looking forward to exploring both fields more deeply and am very thankful to be supported by the NSF GRFP fellowship.” – Nathan Cheng

Alan Chung on Probability and Machine Learning Theory:

“I am interested in problems in Probability Theory and Machine Learning Theory. When applying for the NSF grant, I was working on a project focused on machine learning on graphs. The resulting completed paper, titled Statistical Guarantees for Link Prediction using Graph Neural Networks and co-authored with Professors Amin Saberi and Morgane Austern, was later submitted. In the future, my research interests include more problems in ML theory, as well as topics in probability, such as Random Matrix Theory and Free Probability.” – Alan Chung

Matthew Esmaili Mallory on High-Dimensional Universality with Dependence:

“I am working on universality in the presence of complex dependence structures and high dimensionality, particularly in the case of data augmentation, a machine learning heuristic that synthetically increases the size of training data. Employing data augmentation in a high dimensional setting provides us with quite unique problems, as many classical conjectures about the behavior or convergence of estimators and their risk often end up being untrue. I love working on problems like this which use tools from both pure mathematics and statistics, such as random matrix theory, functional analysis, and ergodic theory." – Matthew Esmaili Mallory

Skyler Wu on Bayesian Identification, Reconstruction, and Forecasting of Dynamical Systems:

“The objective of my project is to develop accurate, numerically stable, and principled Bayesian computational methods for inference of dynamical systems’ parameters, with a demonstrated application to disease forecasting. I have proposed a three-pronged research plan aimed at characterizing the original MAGI [MAnifold-constrained Gaussian process Inference) method’s performance (Yang et al. 2021, developed in this department!), harnessing its inference capabilities to develop innovative disease forecasting models, and improving its computational efficiency and numerical stability. I hope to continue my work under the mentorship of Professor Samuel Kou (Harvard Department of Statistics) and Professor Shihao Yang (Georgia Tech ISyE).” – Skyler Wu

From reviewing these diverse research interests and plans, it’s clear that our students have made a successful, compelling case for receiving the NSF’s GRPF. We congratulate them on this accomplishment and look forward to checking in with them on their projects in the future!