I just got back from a short trip to Washington D.C., where I was attending a partners/planning meeting for the CLiMB 2 project. CLiMB is an acronym that stands for Computational Linguistics in Metadata Building, a fascinating project exploring the use of computational linguistic techniques to partially automate the process of cataloguing scholarly image collections.

This is a very good idea, because museums, libraries, archives and visual resource collections are facing an explosion in the number of digital images that they need to manage, but without a corresponding explosion in resources to adequately catalogue them — and without good cataloguing, it will be more difficult for curators and end users to find the images that they need.

The CLiMB toolkit works by locating references to target objects in digitized scholarly texts (catalogues are ideal, but other texts can be used too), determining the limits of the relevant section of the text, and then using something called a “chunker” to extract noun phrases from the text. These noun phrases can be optionally matched against controlled vocabularies and thesauri, which is another very nice feature, and after being reviewed by a human cataloguer, can be added to the catalogue record for the target object as useful subject terms that will facilitate end user retrieval.

The first phase of the CLiMB project resulted in a prototype toolkit with a relatively user-friendly web-based front end that demonstrated the feasibility of the approach, whereas the second phase of the project is looking to both refine and extend the functionality of the toolkit, and demonstrate the utility of the toolkit in real-life cataloguing scenarios.

I enjoyed the meeting very much; in addition to being an interesting project in its own right, the Primary Investigator Judith Klavans assembled an excellent group of people, so it was also a great networking opportunity for people “in transition” like myself!

Plus I got to hang out in Washington D.C. in decidely spring-like weather with old and new friends, drink Chimay and listen to live jazz at the Eighteenth Street Lounge (former home of Theodore Roosevelt and current home of the Thievery Corporation), and check out the Modigliani exhibition at the Phillips Collection.