Yale University (Yale Univ. Grant & Contract Admin.)

Yale University will investigate how topic modeling can be used to improve search and discovery for users of digital collections. Digital collections may contain hundreds of thousands of images. With limited descriptions associated with these images, Yale is seeking automated ways to connect users to the information they need. The project team will apply topic modeling to three important classes of digital library resources: full-text books, images, and tagged objects. They will build prototypes of user interface applications that will implement topics and test the prototypes to assess the value of topic modeling for users. They will also develop software and tools for topic modeling, which will be made available on the project Web site and in a workshop that will be developed and delivered on how to integrate topic modeling into digital libraries.