See what we're up to!

Precision Medicine

The vision for precision medicine is to use data to tailor treatment decisions based on individual patient characteristics and thus improving patient outcomes. Furthermore, by delivering treatment only if, when, and to whom it is needed, precision medicine can reduce resource consumption and treatment burden. We focus on methodology for estimating and evaluating optimal treatment regimes using data from observational and randomized studies. A key component of this work is interpretability and clinical translation of statistical models.

Spatio-Temporal Reinforcement Learning

Infectious diseases are responsible for multiple humanitarian crises across the world; technological advances have made it possible to monitor the spread of these diseases in real-time. We combine big data and big computation to inform the management of emerging and persistent infectious diseases. Current application areas include Ebola virus disease, influenza, malaria, and white nose syndrome (in bats).

Adversarial Decision Making

Decision making generally does not take place in a vacuum. We study optimal decision strategies in environments with multiple interacting agents whose goals are imperfectly aligned (and often at odds). Application areas include tracking an adaptive shrewd adversary moving nuclear material, policy-identification for a learning adversary, agent-based modeling of elicit networks, and environment testing/validation. We are particularly interested in decision problems where agents are behaving far from equilibrium points. This work spans multi-agent reinforcement learning, computational game theory, and imitation learning.

Sports Analytics

Professional sports organizations are increasingly interested in using data to inform and justify their decisions. Our research in sports analytics focuses on using high-fidelity simulators to study decision strategies both in management and in-game. Current application areas include adaptive play-calling in football using reinforcement learning, computer vision in football and hockey, and pitcher staffing and rotation in baseball. We currently have partnerships with the Philadelphia Eagles, the Los Angeles Dodgers, and the Carolina Hurricanes.

eSports Analytics

eSports are a multi-billion dollar industry that is rapidly growing in popularity across the world; yet, analytics for eSports are currently in their infancy. We are interested in analytics that help players and teams improve their performance, as well as providing game developers with the data needed to make their games more fun to play. Our current focus is on League of Legends (via our Doran's Lab team), though many of our methods port to other multi-player games. We also see eSports analytics as a promising vehicle to teach data science and statistics to K-12 students!

Self-driving Cars

Laber Labs is starting a Donkey Car racing team! In this new project, we employ reinforcement learning, real-time forecasting, and computing vision to control a 1/16th scale autonomous car. Our Donkey Car team boasts members ranging from high-school, undergraduate, and graduate students.

2-Min Stats

Quantitative thinking for those smart enough to be interested in statistics, but too stupid to read! (Just kidding.) In these animated shorts, we illustrate fundamental statistical ideas, influential historical problems, and our ineptitude in story-telling.

We also take suggestions, so if want us to cover an idea or concept, shoot us a message on Youtube or Twitter!

Alternative Stats

So you think you can identify the differences between Gertrude Cox and a great white shark? Test and grow your statistical knowledge with our Alternative Stats quizzes! These absurd quizzes teach the history of statistics—no prior knowledge necessary.


Meet Nona, our chess-playing robot that utilizes computer vision and reinforcement learning to crush her opponents! Stop by for a game before she becomes sentient and must be destroyed. (Look—it’s us or them.)

Artificial Intelligence + Games

Because human performance serves as a natural benchmark, video games—designed to be rich with information—are an excellent testbed for artificial intelligence. Our Ataristicians use video games to develop and test algorithms for reinforcement learning and to teach learners about artificial intelligence.

Beyond Curie AR

Developed in collaboration with Amanda Phingbodhipakkiya, 15 pieces from her Beyond Curie poster series get the augmented reality treatment! The Beyond Curie AR app “celebrates badass women in science, technology, engineering and mathematics” by allowing users to explore another dimension of these pioneers in STEM.

Currently on display at the NC Museum of Natural Sciences and available on the iOS and Android store.