| Content analysis and unstructured information management |
With the explosion of the number of information items that each person generates and uses in their professional and personal lives, the need for better techniques for managing those large collections of loosely related items is becoming a top priority. We are pursuing new methods for content analysis, for example in document and image similarity. We are also investigating methods for obligation and policy management for documents and files based on the content and metadata associated with those items.
Our research in this area focuses on feature selection and extraction, scalable clustering algorithms, locality-sensitive hashing, distributed metadata management, efficient data scanning, and privacy-preserving minimally invasive policy mechanisms.
Selected Publications:
|  |
|
Printable version
