A faculty member in the School of Information Studies at Syracuse University is using artificial intelligence to improve the National Science Foundation’s grant reviewing process.

Daniel Acuna, an assistant professor in the iSchool, was recently awarded a two-year grant by the NSF with nearly $170,000 for which he is the principal investigator. The grant is allocated for improving grant reviewing and strengthening scientific innovation through linking funding and scholarly literature.

Scientists and researchers need extraordinary ideas when applying for a grant opportunity from the NSF, Acuna said. But grant writers and researchers often have difficulty knowing what fields and topics have yet to be covered and funded.

“Everybody thinks that finding grants and finding information about funding from federal agencies is like Googling,” Acuna said. “Finding novel ideas with the current tools is really, really hard.”

Acuna said a lot of funding agencies tend to choose novel ideas, so he wanted to design a system that would help scientists make better decisions. He said it would also help program officers at the NSF pick better applicants and determine who to fund.

Currently, Acuna said, scientists have to go through hundreds of grants and try to understand what those grants mean regarding a relationship between publications, citations and grants. To solve this problem, Acuna said he wants to use artificial intelligence to systematize this idea of analyzing science.

In this way, scientists can begin to find patterns in the data and have better ways to produce knowledge and look for funding ideas, Acuna said.

“Every year there are millions of publications (and) no one scientist alone can read those publications by themselves,” Acuna said. “So we need an AI-based system to pre-process those datasets and present with what we want.”

He said AI will help a lot in the process by automatically combining the contents of the publication and grants to present the results to the scientists.

AI presents a lot of opportunities for scientists and can help them make more accurate decisions in matching contents, Acuna said. This makes it faster to explore the entire set of publications and grants.

Acuna added that he plans to work on the project for two years. It will focus on gathering and pre-processing the data in the first year and developing web-based tools that will allow scientists and program officers to navigate the data in the second year.

“Relatively soon, a lot of things scientists do could be improved by using AI or computers,” Acuna said. “That is a very positive thing because we can make science more efficient and get resources when we really need them.”

Konrad Kording, a professor and a former colleague of Acuna at Northwestern University, said if the idea is successfully implemented, scientists or researchers can have better scientific decision making, better funding for important projects and less waste of resources.

Acuna will be collaborating with Kording, Associate Professor Bei Yu at the iSchool and James Evans at University of Chicago.

Everything they produce, including software, datasets and analyses will be available to the public, Acuna said.

“I’m very excited for that possibility and positive about the prospect of using AI,” Acuna said.

Published on October 5, 2016 at 11:47 pm

Contact Chieh: chchen@syr.edu