Judith Klavans
Director of Research, Center for Advanced Study of Language
University of Maryland
Marilyn Domas White
Associate Professor
University of Maryland
Angela Giral
Director (retired), Avery Library
Columbia University
The goal of Phase One of the CLiMB (Computational Linguistics for Metadata Building) project was to discover to what extent and under which circumstances computational linguistics techniques could be used to extract descriptive metadata from texts associated with image collections. Toward that end, CLiMB developed a prototype toolkit, which enabled the identification and extraction of terms from text describing images. Criteria for selecting collections were established, and evaluation of the toolkit was performed over a set of initial collections. Originally funded at Columbia University, CLiMB has received a planning grant for Phase Two, in which the project is transitioning to its new home at the University of Maryland (UMD). The goals of “CLiMB-2” are to:
• identify a set of collections appropriate for testing CLiMB
• improve the CLiMB toolkit with the goal of helping the cataloger work more efficiently (i.e. more breadth in less time)
• apply the improved toolkit to extract descriptive metadata for these collections
• study the usability of the CLiMB toolkit and evaluate how it helps users
We are currently identifying museum and library partners with appropriate collections and user groups with a view toward evaluating the modified toolkit and assessing its impact on retrieval in real settings. The purpose of this project briefing is to present the vision and implementation plan for CLiMB-2 at UMD and to build additional relevant partnerships.
http://www.columbia.edu/cu/libraries/inside/projects/climb
http://www.umiacs.umd.edu/~climb
Handout (MS Word)
