Dutta to receive NSF CAREER Award

ECE Faculty Member Sanghamitra Dutta wins the NSF CAREER Award for her project titled, “Information-Theoretic Measures for Fairness and Explainability in High-Stakes Applications.” This highly competitive award, considered one of NSF's most prestigious awards in support of early-career faculty, is given to researchers who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization. Dutta will be awarded $665,552 over five years from the NSF Communications & Information Foundations (CIF) Program.

Dutta’s project seeks to advance the foundations of ethical and socially-responsible machine learning by empowering users to systematically identify, explain, and mitigate the sources of disparity. Rethinking the traditional paradigm of separately addressing fairness and explainability, this research project will jointly examine fairness and explainability through a unified information-theoretic lens.

The project will provide a novel information-theoretic view of responsible machine learning, by leveraging a body of work in information theory called Partial Information Decomposition (PID). Four research thrusts will be investigated: (i) Providing an information-theoretic framework for explaining sources of disparity with respect to protected attributes such as gender, race, etc.; (ii) Performing systematic feature selection and representation learning with disparity control; (iii) Investigating fundamental limits with a focus on distributed and federated settings; and (iv) Validating these findings on real-world datasets in finance and education.

Furthermore, through extensive outreach and student engagements on the social impacts of machine learning, this project aims to instill interest in mathematically-principled approaches and STEM education among students to spearhead the next generation of socially-responsible technology.

A member of the UMD ECE faculty since 2022, Dutta has published actively in several leading machine learning conferences. Her primary research interests lie in reliable and trustworthy artificial intelligence for social good. A recent paper from Dutta’s research group on explainability, titled “Robust Counterfactual Explanations for Neural Networks with Probabilistic Guarantees”, was presented at the 2023 International Conference on Machine Learning (ICML).

Another recent work from her group titled, “Demystifying Local and Global Fairness Trade-offs in Federated Learning Using Partial Information Decomposition,” will be presented at the upcoming 2024 International Conference on Learning Representations (ICLR). This work has also featured in the Montreal AI Ethics Brief.

Published January 18, 2024