Photo Credit: Philip Fei Wu
College of Information Studies
Hornbake Library, South Wing
University of Maryland
College Park, MD 20742-4345
Telephone: (301) 405-2033
Judith L. Klavans is the Principal Investigator on the Mellon-funded Computational Linguistics for Metadata Building (CLiMB) research project, now based at the College of Information Studies at the University of Maryland. In addition to leading the project, she is involved in developing analysis and filtering techniques for the extraction of metadata, particularly through thesaurus-driven disambiguation. She is also involved in the Defense Advanced Research Projects Agency (DARPA)-funded TIDES multilingual multimedia summarization project, in which her primary technical role is the areas of utility evaluation and in coherence for summarization. She is currently Research Professor at the College of Information Studies at UMD
Klavans holds a Ph.D in Linguistics from the University of London and has worked on numerous computer science, digital library and digital government projects. In particular, she has served as principal investigator on several other large research projects, including the National Science Foundation (NSF)-funded PERSIVAL medical digital library, the NSF and Bureau of Labor Statistics (BLS) supported Digital Government Research Center joint project with University of Southern California-ISI, and the Defense Advanced Research Projects Agency (DARPA)-funded TIDES multilingual summarization project. Her research interests include linguistics, digital library research, language, and natural language systems.
Klavans initiated the CLiMB project at Columbia University in 2002. From the beginning, the goal of this project has been to explore ways to apply computational linguistic techniques over scholarly texts as a means of extracting metadata to populate image catalog records. Since its inception at the University of Maryland, Klavans has led the CLiMB team in extending the Toolkit’s overall functionality and improving its effectiveness through disambiguation techniques and the integration of controlled vocabularies and metadata schemas. Through these and other cutting edge approaches, the CLiMB Toolkit has the potential to greatly improve consistency and accessibility across collections.