My interests lie in the areas of databases, data mining, information retrieval, and the semantic web, as they apply to multimedia information.

My previous work was in content-based retrieval and focused on showing how information from several modalities could be used to improve retrieval over a single modality. Specifically, my students and I have shown how text information can be used to improve image-based searches and how images can be used to improve text-based searches.

We are now using this work to develop a social-based theory of web semantics, and techniques for multimedia annotation. These techniques rely heavily on various dimensional reduction techniques, such as latent semantic analysis (LSA). As such, we are also examining these techniques in more detail, trying to determine the actual semantics of the various dimensions in the reduced space.