CS Goes to the White House: NSF-led National AI Research Resource (NAIRR) Pilot Awards Access to Stony Brook Researchers

Stony Brook Researcher Harsh Trivedi presents at the White House
Stony Brook Researcher Harsh Trivedi presents at the White House

The National Science Foundation (NSF) and the Department of Energy (DOE) have awarded access to advanced computing resources to 35 projects through the National Artificial Intelligence Research Resource (NAIRR) Pilot. Prof. Niranjan Balasubramanian from the Department of Computer Science (CS) at Stony Brook University is one of the recipients of the NAIRR program.

The NAIRR initiative aims to democratize access to AI research tools across the United States. The selected projects span critical areas such as deepfake detection, AI safety, and medical diagnoses, utilizing high-performance computing systems from various NSF-funded and DOE-supported facilities, including Delta, Frontera, Lonestar, and the AI Testbed.

On May 6th, Stony Brook University doctoral researcher Harsh Trivedi represented CS at the White House Office of Science and Technology Policy (OSTP) launch event to celebrate the first allocation of the NAIRR Pilot. At the event, researchers presented overviews of their projects, highlighting the critical work supported by the NAIRR Pilot. This gathering underscored the government’s commitment to advancing responsible AI research and innovation.

Balasubramanian’s approved project, Developing Reliable Agents for Automating Complex Digital Tasks, aims to develop Large Language Model (LLM) agents for automating complex digital tasks via interactive coding and API calls. The project builds on his lab’s past work, called AppWorld, which is an execution environment of simulated real-world apps (e.g., Amazon, Spotify) operable via APIs, and an associated benchmark of day-to-day tasks posed on these apps. The NAIRR project will train LLM agents using direct and self-supervision strategies on this environment and benchmark. This project exemplifies the innovative research being enabled by the NAIRR Pilot.

To learn more about AppWorld, please visit https://appworld.dev. To learn more about specific NAIRR projects or to apply for access to resources, please visit the NAIRR portal at https://nairrpilot.org/

UPDATE: AppWorld won the ACL 2024 NLP Best Resource Paper Award! Congratulations to Harsh and the team!

-Sahil Sarna