School of Information Studies’ researchers provide virtual bricks and mortar for the National Science Digital Library
School of Information Studies’ researchers provide virtual bricks and mortar for the National Science Digital LibraryJune 30, 2003Judy Holmesjlholmes@syr.edu
When the National Science Digital Library (NSDL) opened its virtual doors last December, researchers from the School of Information Studies helped turn the key. Sponsored by the National Science Foundation (NSF), the ambitious project aims to be the most comprehensive online source for science, technology, engineering, and mathematics education in the country and is emerging as a center of innovation in digital libraries as applied to education.
Researchers in the school’s Information Institute of Syracuse (IIS) and the Center for Natural Language Processing (CNLP) have been leaders in providing ideas and innovative virtual bricks and mortar to build the NSDL (www.nsdl.org). Some of those bricks and mortar were on display during the NSDL’s official launch in Washington, D.C., on Dec. 3, 2002, and include the NSDL’s GEM Subject Browser, AskNSDL, Bill Nye the Science Guy’s Information Safari, and a direct link to the Virtual Reference Desk on the library’s “Help” page.
“The role we are playing in the creation of the NSDL is a validation of the impact we are having on the lives of students, educators, and the research community,” says R. David Lankes, executive director of the Information Institute of Syracuse, who is chair of the NSDL technology committee. “We entered into this digital library community with no special advantage and have quickly assumed a central leadership role not just for our metadata work, our digital reference work, ERIC/IT, or AskERIC, but for our entire approach to education information. The recognition we have received from the NSF for our work is a significant honor for the entire institute.”
Since the NSF issued its first “call for proposals” in 2000, researchers in the IIS and the CNLP, which is directed by Prof. Elizabeth Liddy, have garnered more than $2.2 million in NSDL research grants. The projects range from developing automatic metadata generating systems for the library’s collections to developing the NSDL’s digital reference gateway. All of the research grants build on the expertise and technology already developed by the IIS and the CNLP in the areas of digital reference, metadata tagging, and natural language processing.
“The NSF uses an analogy that refers to the NSDL as a ‘national treasure’ that anyone can access for math, science, and technology information,” Lankes says. “Our researchers are among the primary architects who are building the crucial components of this national treasure.”
The virtual bricks and mortar supplied by the SU research teams resulted from several grants:
- a $490,000 grant awarded to the CNLP in partnership with the University of Washington and solutions-united of Syracuse to use natural language processing technology to develop a faster method of tagging and cataloging NSDL materials; a $475,000 grant, in collaboration with the University of Washington, to extend that capability by developing a middleware tool that will automatically assign content standards and benchmarks to NSDL educational materials;
- The CNLP has also received a $375,000 grant to evaluate the benefits of catalog information to users within the search and resource selection process, and a $100,000 grant to integrate the center’s automatic metadata generation capabilities in the NSDL for use by all collection providers so that the collections can grow more quickly by automating the metadata assignment task;
- a $350,000 grant awarded to the IIS in collaboration with researchers from Ohio State University and the University of Illinois to build the NSDL’s virtual reference gateway; and
- a $450,000 grant awarded to the IIS in partnership with the University of Massachusetts, Amherst, and the Maryland-based Wondir Foundation, to merge the information retrieval and digital reference components of the NSDL so that users can find answers to questions regardless of whether the answers come from the documents in the NSDL collection or experts accessible through the NSDL’s virtual reference desk.
Significant progress has been made in all the research areas. “The CNLP is continuing its research to automatically assign catalog and educational content standards to educational activities and lesson plans within the NSDL collection,” Liddy says. “In addition, the center is using a new $100,000 grant to incorporate its metadata generating technology directly into the NSDL for use by collection providers who do not have the means to produce catalog information.”
The CNLP is also working jointly with Cornell University’s Human Computer Interaction laboratory on an evaluation project called “MetaTest.” The results of the evaluation research will enable the NSF to understand the value, effectiveness, and efficiency of metadata, whether manually or automatically assigned, Liddy says. “The research will also enable the NSF to understand, for the first time, how NSDL patrons actually use metadata in their information seeking.” MetaTest is supported by a $375,000 grant to the CNLP.
Under its initial $350,000 grant, the IIS significantly improved and enhanced its digital reference software tool and built the NSDL’s virtual reference service called AskNSDL, says Joanne Silverstein, IIS associate director. AskNSDL went live in December when the initial NSDL portal was launched.
“AskNSDL provides students, teachers, parents, and other users with a simple, user-friendly way to connect with the researchers, scientists, and librarians who are building the portal and organizing the content that it provides access to,” Silverstein says. “In addition, AskNSDL can capture the answers that are provided and make the answers available to other users who ask similar questions.”
Work is continuing on the $450,000 project that is designed to scale up AskNSDL so that it can serve more and more users. In cooperation with the Center for Intelligent Information Retrieval (CIIR) at the University of Massachusetts, Amherst, and the Wondir Foundation, the IIS is investigating ways to automatically identify and separate out questions submitted by users that could potentially be answered by an automated system, rather than needing human intermediation.
“Researchers are currently analyzing thousands of previously submitted digital reference questions,” Silverstein says. “The plan is to combine machine-learning techniques developed by the CIIR, digital reference expertise from the IIS, and current work on faceted classification schemes for digital reference questions to build an automated system that could be implemented with AskNSDL.”