The Harvard Data Science Initiative (HDSI) has received a generous grant from the Alfred P. Sloan Foundation to support a team of Harvard faculty working on new methods for causal inference and machine learning.
“Existing causal inference methods largely address simple situations where the intervention is binary—for example, receiving treatment or not—and the observations are independent across time and space,” explains Francesca Dominici, Clarence James Gamble Professor of Biostatistics, Population, and Data Science at the Harvard T.H. Chan School of Public Health, HDSI Faculty Co-Director, and the project’s Principal Investigator. “But observational data from the real world is never this simple. When assessing the causal effects, interventions are often complex, they involve multiple actions simultaneously, and they are measured on a continuous scale. Data scientists require new methods and software tools that can meet the challenge of identifying cause when data is imperfect, messy, and influenced by many factors over time.”
The grant, totaling about $1 million, enables three collaborative projects led by faculty, including Dr. Dominici who will lead the design of new methods for characterizing subpopulations that experience different causal effects of a given intervention.
The HDSI will serve as a convener and coordinator for the project, including creating outlets for the greater data science community to access emerging results. “Uniting and amplifying these types of methodological advances across varied application domains is precisely what the HDSI aims to achieve,” Dominici shares. “We’re incredibly grateful for the Sloan Foundation’s investment in this work.”
