research interests

I’m broadly interested in statistics, machine learning, and optimization. I’m also interested in lots of applications: to finance, operations research, public policy, social good, sustainability, epidemiology, healthcare, autonomous vehicles, analytics, …. A lot of my recent work has focused on building reliable and trustworthy machine learning systems, by taking a close look at the (many) statistical and computational issues that arise after a machine learning model has been deployed into real-world systems and scientific applications.

On a more technical level, I’m interested in – and have recently worked on projects related to – the following areas:

• distribution shift, robust optimization, subpopulation-level performance, responsible AI
• conformal inference, distribution-free uncertainty quantification
• weak supervision
• tuning parameter-free stochastic optimization
• implicit regularization
• sparse regression
• large-scale multiple testing
• risk estimation, model selection
  
  

representative publications

Here are a few (representative) publications related to the above topics:

The Lifecycle of a Statistical Model: Model Failure Detection, Identification, and Refitting

We develop methodology (and associated theory) that (i) tracks deployed model performance, (ii) raises an alarm when performance degrades, and (iii) generates interpretable diagnostics.

Robust Validation: Confident Predictions Even When Distributions Shift

We develop methodology and theory that repairs and protects any deployed model from future degradations in performance without retraining.

Minimizing Oracle-Structured Composite Functions

We develop reliable algorithms for (re)training models without carefully tuning step sizes.

A Continuous-Time View of Early Stopping for Least Squares

We develop statistical theory advancing our understanding of popular optimization heuristics.

all publications