Travel Grant: Reinforcement Learning: from Theorem to Real-World

This grant provides funding for a workshop entitled "Reinforcement Learning: from Theorem to Real-World" to be held in Boston, Massachusetts, 23-24 January 2025. Despite Reinforcement Learning (RL) growing in influence across fields like healthcare, robotics, energy systems, and transportation, a substantial gap remains between theoretical progress and practical deployment.

This workshop on RL aims to close this gap by uniting experts from computer science, control theory, machine learning, and specific application areas to focus on core issues, including aligning RL theory with practical requirements, incorporating domain-specific structures into RL algorithms, applying RL in diverse real-world settings, and refining benchmarks and performance metrics. The ultimate objective is to create a forward-looking RL research agenda that promotes both theoretical rigor and practical feasibility.

This workshop will have broad impacts by fostering interdisciplinary collaboration and encouraging innovations that can improve applications across fields. Additionally, it will focus on education and training strategies to broaden participation in RL, preparing students from varied disciplines and backgrounds to work in this growing area.

The intellectual contributions of the workshop are three-fold: (1) it will convene a complementary group of RL experts to collaboratively address common technical challenges, establishing a focused research agenda for RL; (2) it will identify critical gaps in theory, algorithm design, dataset availability, and benchmarking practices, enabling advancements that enhance RL’s reliability and utility across applications; and (3) it will enrich RL research through interdisciplinary insights from engineering, societal systems, science, and computational sciences to foster a more holistic approach to RL development. By gathering researchers across disciplines, the workshop aims to establish a research, education, and application framework for RL that is adaptable, rigorous, and inclusive.

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.