University of Hawaii, Hilo contributes significantly to a large-scale AI initiative
University of Hawai'i at Hilo Joins $152 Million Artificial Intelligence Project
The University of Hawai'i at Hilo (UH Hilo) is set to play a significant role in a groundbreaking five-year project led by the Allen Institute for Artificial Intelligence (Ai2). The project, known as Open Multimodal AI Infrastructure to Accelerate Science (OMAI), aims to develop the first fully open-source, multimodal large language models (LLMs) trained specifically on scientific data and literature.
The OMAI project has secured a combined $152 million in funding from the National Science Foundation (NSF) and AI company Nvidia. The initiative addresses the high costs of advanced AI development that many academic labs and federally funded researchers face, thereby democratizing access to cutting-edge AI infrastructure for the broader scientific community in the U.S.
UH Hilo has been awarded over $1.4 million as a key collaborating institution in this project. The university's involvement will focus on integrating state-of-the-art AI infrastructure into its academic programs, including data science and computer science. This includes creating summer internships, school-year research opportunities, and community engagement efforts to ensure the AI models address diverse scientific needs, especially those relevant to Hawai'i's local researchers.
Dr. Travis Mandel, associate professor and data science program coordinator at UH Hilo, serves as co-principal investigator and co-leads the adaptation and community engagement components to tailor the AI technology for different scientific communities. His involvement also emphasizes increasing transparency in AI and educating students and researchers on AI's inner workings and utility in solving real-world problems.
The OMAI project is part of the White House AI Action Plan to bolster U.S. leadership in AI-driven scientific discovery and democratize access to essential AI software infrastructure. The project aims to create AI for good and science, as stated by Mandel.
| Aspect | Details | |----------------------------|-------------------------------------------------------------------------------------------------| | Project Name | Open Multimodal AI Infrastructure to Accelerate Science (OMAI) | | Funding | $152 million total: $75M NSF + $77M Nvidia | | Purpose | Develop open-source multimodal AI models trained on scientific data to accelerate research | | Lead Organization | Allen Institute for AI (Ai2) | | Key Partners | Multiple universities including University of Washington and University of Hawai'i at Hilo | | UH Hilo Involvement | $1.4M funding; AI infrastructure integration in academic programs; student internships/research | | Focus at UH Hilo | Data Science and Computer Science programs; community engagement; AI model adaptation | | Technology Provided | Nvidia HGX B300 systems and AI hardware infrastructure |
Mandel's stance is that AI for science is important, despite the energy and water usage it requires. Specialists will visit UH Hilo for yearly scheduled workshops. UH Hilo students will contribute to the development of AI models remotely to benefit their education. Mandel and other team members will make recurring trips to Seattle. The project will also allow students to work with Ai2 professionals remotely.
At present, there are no subscription, login, or registration details associated with this project. More information about the OMAI project and its impact on UH Hilo and the broader scientific community will be shared as the project progresses.
- The University of Hawaii at Hilo's involvement in the OMAI project involves integrating advanced technology like AI infrastructure into its academic programs such as data science and computer science, contributing to the education of its students.
- Dr. Travis Mandel, a key figure in the OMAI project, emphasizes the importance of AI for scientific advancements, recognizing that it requires resources from the environment like energy and water, while also striving to make AI more transparent and acceptable for diverse scientific needs.
