Epscor research fellows: NSF: building a large predictive model for severe flooding events

The National Science Foundation Office of Integrative Activities awarded the University of Louisville $299,676 on January 1, 2027, under the Integrative Activities program (CFDA 47.083) to develop a large predictive model for severe flooding events through the EPSCOR Research Fellows initiative.

The award funds a fellowship for an assistant professor at the University of Louisville and training for a graduate student to advance development of a framework for predictive common operating picture synthesis in flood crisis response. The project leverages artificial intelligence foundation models to unify fragmented flood-relevant data sources, predict near-term impact evolution, and synthesize information for reliable decision-making in flood emergency response. The work develops a reinforcement-learning-based orchestration mechanism that enables the system to learn how to select, coordinate, and use specialized model capabilities according to analytical needs rather than relying on a single model or fixed processing pipeline. The research aims to enhance disaster situational awareness, enabling more effective resource allocation, faster response times, and reduction in loss of life and economic impacts caused by flood disasters.

The University of Louisville Office of Research & Innovation serves as the recipient. Work is performed at Oak Ridge, Tennessee in collaboration with the Environmental Sciences Division at Oak Ridge National Laboratory. The period of performance runs from January 1, 2027, through December 31, 2028. The project strengthens research infrastructure at the University of Louisville by supporting faculty professional advancement and establishing long-term research collaboration between the principal investigator and Oak Ridge National Laboratory, with project outcomes to be incorporated into courses taught by the principal investigator.

Abstract

This research infrastructure improvement (RII) epscor research fellows project provides a fellowship to an assistant professor and training for a graduate student at the university of louisville (uofl). This work is conducted in collaboration with the environmental sciences division at OAK ridge national laboratory (ornl). Through the fellowship, the PI will advance the development of a framework for predictive common operating picture synthesis in flood crisis. The project will leverage and synthesize complementary capabilities of existing artificial intelligence (AI) foundation models and tackle the challenge of coordinating them for flood-related predictive intelligence. The project results will enhance disaster situational awareness, enabling more effective resource allocation, faster response times, and ultimately, a reduction in the loss of life and economic impacts caused by flood disasters. This project will harness multiple AI foundation models and flood-relevant data sources to unify fragmented evidence, predict likely near-term impact evolution, and synthesize information for a reliable decision-making in flood emergency response. Particularly, IT will develop a reinforcement-learning-based orchestration mechanism for complementary foundation models, enabling the system to learn how to select, coordinate, and use specialized model capabilities according to the analytical needs instead of relying on a single model or a fixed processing pipeline. This research will strengthen research infrastructure at uofl by supporting faculty professional advancement in the intersection of flood crisis and advanced AI techniques and training one graduate student. In addition, the project will establish research collaboration between the PI and the ornl for their long-term career development. Furthermore, project outcomes will be incorporated into courses taughted by the PI for curriculum enhancement. Finally, IT will promote scientific dissemination and outreach within the broader research community. This project is supported by the epscor research infrastructure improvement program: epscor research fellows (ERF). The ERF program supports early- and mid-career investigators in eligible jurisdictions to develop collaborations at the nation?s private, government or academic research institutions. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.