CSAIL’s Broderick and Carbin earn Presidential Early Career Awards

MIT EECS associate professors and CSAIL principal investigators Tamara Broderick and Michael Carbin.

Earlier this month, Tamara Broderick and Michael Carbin, both associate professors in MIT’s Electrical Engineering and Computer Science Department (EECS) and CSAIL principal investigators, received the Presidential Early Career Award for Scientists and Engineers (PECASE). Broderick received her honor from the Office of Naval Research for her work on robust machine learning, while Carbin received his award from the U.S. government via the National Science Foundation for his research on programming systems.

Former President Joe Biden announced this honor amongst nearly 400 other recipients, including CSAIL principal investigator Christina Delimitrou. Established in 1996, the PECASE recognizes scientists and engineers who “show exceptional potential for leadership early in their research careers.”

Tamara Broderick: Quantifying uncertainty amidst noisy data

Recently tenured at MIT, Broderick is also a member of the MIT Laboratory for Information and Decision Systems (LIDS), the MIT Statistics and Data Science Center, and the Institute for Data, Systems, and Society (IDSS).

She now aims to help people understand the limits of data analysis tools and how they can be improved. Her work includes providing fast and accurate quantification of uncertainty and robustness for scientific conclusions. Some of her work innovates in the area of Bayesian inference — a mathematical approach where you can continuously update your assumptions as more data comes in. For instance, Bayesian inference is used to predict the winners of elections using polling data.

Broderick and collaborators recently developed an optimization method to speed up Bayesian inference, achieving more accurate results while requiring less work from the user. The automated technique allows scientists to input their model; the scientist can then get estimates and uncertainties for things like the impact of microcredit loans in developing nations or top athletes in a particular sport.

The research group she leads also worked with collaborators to develop a machine-learning model that can predict ocean currents more accurately than conventional approaches. It uses knowledge from fluid dynamics to model ocean current physics more realistically while requiring minimal additional computational expenses compared to standard methods.

Her previous honors include designation as an IMS Fellow, selection to the COPSS Leadership Academy, a Susie Bayarri Lecture, an Early Career Grant from the Office of Naval Research, an NSF CAREER Award, a Sloan Research Fellowship, an Army Research Office Young Investigator Program award, Google Faculty Research Awards, an Amazon Research Award, the Berkeley Fellowship, an NSF Graduate Research Fellowship, a Marshall Scholarship, and the Phi Beta Kappa Prize.

Michael Carbin: Manipulating uncertainty to improve system efficiency

Carbin previously received his master’s degree (2009) and PhD (2015) at MIT. Now, he leads the Programming Systems Group at CSAIL. 

Currently, he’s working on programming systems that manipulate uncertainty to improve performance, energy consumption, and resilience. His group refines the design, semantics, and implementation of these systems.

For instance, his group recently explained why it’s so challenging to implement a quantum algorithm to run on a quantum computer using traditional, high-level programming abstractions. His team presented a model for a more user-friendly quantum computer that offers the ability for higher-level programming while still noting that a quantum algorithm needs all instructions to be reversible to process quantum information correctly.

Another example: His team and their collaborators just introduced a programming interface called “inference plans for probabilistic programming.” Probabilistic programming is a methodology with supporting programming languages for writing programs that must soundly model and act on uncertainty in their environment — such as the noise in a radar sensor.  The new interface enables developers to soundly navigate the trade-offs between precise and approximate probabilistic inference to produce an accurate and performant probabilistic program.

The PECASE is one of many honors in Carbin’s career, which includes the NSF CAREER Award, the Sloan Research Fellowship, CRA Skip Ellis Early Career Award, the MIT Frank E. Perkins Award for Excellence in Graduate Advising, the MIT Louis D. Smullin Award for Teaching Excellence, a Facebook Research Award, a Google Faculty Research Award, the Microsoft Research Graduate Fellowship, the MIT Lemelson Presidential Fellowship, and several best paper awards.