Giles and Mitra make strong showing at JDCL 2011

July 28, 2011

Since 2001, the Joint Conference on Digital Libraries (JCDL) has served as the major international forum focused on digital libraries and associated technical, practical and social issues. At this year’s conference, which was held June 13-17 in Ottawa, Canada, Penn State researchers made a strong showing. The theme for JCDL 2011 was "Digital Libraries: Bringing Together Scholars, Scholarship and Research Data,” in recognition of the changes the digital age is now bringing to scholarship.

Two professors in Penn State’s College of Information Sciences and Technology, C. Lee Giles and Prasenjit Mitra, presented papers at the conference. In keeping with the theme of the conference, Giles, along with his collaborators, introduced a public website that is designed to connect researchers who share similar interests.

Giles, David Reese professor of information sciences and technology, Graduate School professor of computer science and engineering in the College of Engineering, and director of the Intelligent Systems Research Laboratory, presented two long papers: “CollabSeer: A Search Engine for Collaboration Discovery” (co-authored with Hung-Hsuan Chen, Liang Gou and Xiaolong Zhang) and “On Identifying Academic Homepages for Digital Libraries” (co-authored with Sujatha Das, Mitra and Cornelia Caragea); and a short paper, “Ranking Authors in Digital Libraries” (co-authored with Sujatha Das and Mitra).

In addition to the papers he co-authored with Giles, Mitra, associate professor of information sciences and technology, presented “Event Detection with Spatial Latent Dirichlet Allocation” with Chi-Chun Pan, a research assistant in the Department of Industrial and Manufacturing Engineering at Penn State.

At the conference, Giles, along with Chen, a graduate student in Computer Science and Engineering, Gou, a graduate student in IST, and Zhang, an associate professor of information sciences and technology, unveiled CollabSeer, a search engine for discovering potential collaborators for a given author or researcher. CollabSeer discovers collaborators based on the structure of the coauthor network and a user’s research interests.

The project is funded by Alcatel-Lucent USA Inc., a French-owned telecommunications company that was spun off from AT&T Technologies in 1996.

Giles and his collaborators started working on CollabSeer last fall, he said. Previously, due to difficulty in extracting data, there was no search engine for finding potential collaborators. Microsoft Academic Search, a free search engine for academic papers and resources principally in the field of computer science, does not provide collaborator recommendations.

“If you’re looking for collaborators for research projects, (CollabSeer) could be a very useful site,” he said.

CollabSeer, which currently features about 400,000 authors, provides a ranking of collaborators based on needs and research interests. In a similar way that social media websites such as Facebook connect people through friend-of-a-friend networks, Giles said, CollabSeer recommends “collaborators of collaborators.”

CollabSeer is a beta site, which means that it is still in the initial testing stages. Giles and his team plan to make adjustments based on feedback from users and explore different methods of visualizing collaboration networks. While the site primarily covers computer science, he said, it can be applied to other areas.

“It’s very easy to extend this model to other fields,” Giles said.

The other long paper that Giles presented at the JCDL, “On Identifying Academic Homepages for Digital Libraries,” was co-authored by Sujatha Das, a graduate student in Computer Science and Engineering, Mitra, and Cornelia Caragea, a post-doctoral scholar at IST. The paper addresses the questions of how many homepages of researchers are on the web, if it is possible to accurately discriminate between academic homepages and other webpages, and what information can be extracted about researchers from their homepages.

Giles, along with Das and Mitra, presented a short paper at the conference, “Ranking Authors in Digital Libraries.” Searching for people with expertise on a particular topic, also known as expert search, is a common task in digital libraries. Das, Mitra and Giles, in their paper, propose graph-based models that accommodate multiple sources of evidence in a PageRank-like algorithm for ranking experts.

“It’s a way of evaluating the importance or 'celebrity' of work,” Giles said.

In addition to presenting the three papers at JCDL 2011, Giles also co-presented a workshop, “Digital Repositories and Field-Specific Libraries: Opportunities and Challenges.” A disciplinary repository is a collection of works or data associated with these works of scholars in a particular subject area. At the workshop, Giles, along with Salvatore Mele (CERN) and Simeon Warner (Cornell) and workshop participants, shared "secrets for success” and allowed discussions of technology, services, interoperability and the engagement of users.

Last Updated July 28, 2011