Federal Grant Award Summary

Carnegie Mellon University received a $150,000 Project Grant from the National Science Foundation's Division of Mathematical Sciences under the Mathematical and Physical Sciences (CFDA 47.049) program, effective July 1, 2026 through June 30, 2029. This collaborative research initiative will develop statistical and machine learning methods for measuring data value in artificial intelligence model training and data-driven decision systems. The project will produce three primary research deliverables: (1) priority-aware valuation rules that incorporate originality, provenance, and individual risk considerations within a unified axiomatic framework for AI data attribution; (2) efficient estimation and approximation algorithms for Shapley-value-based data valuation in high-dimensional and large-scale learning settings; and (3) population-level theory addressing statistical and computational limits of approximating Shapley values in contemporary AI models.

Beyond core research outputs, the award will support substantial educational and dissemination components, including graduate and undergraduate training, development of educational materials, public dissemination of research results, and open-source software development for the broader AI and data science communities. The project addresses fundamental challenges in creating transparent and fair AI data ecosystems by establishing mechanisms to appropriately credit and compensate data contributors while ensuring robustness against strategic manipulation and enabling principled uncertainty quantification in data valuation assignments.