University of Minnesota

Ravindra Kasturi

Jaideep Srivastava

Ayush Singhal

Computerized systems for automating content annotation (e.g., tag creation and/or expansion) for low-content items within a computer network by leveraging intelligence of other data sources within a network to generate secondary content (e.g., a 'context') for items (e.g., documents) for use in a tagging process. For example, based on user assigned tags for an item, secondary content information can be generated and used to determine a new list of candidate tags for the item. Additionally, the context of an input item may be compared against the respective contexts of a plurality of other items to determine respective levels of similarity between the input item and each of the plurality of other items in order to annotate the input item. Techniques involving web-distance based clustering and leveraging crowd-sourced information sources to remove noisy data from annotated results are also described.

Growing community of inventors

Redmond, WA, United States of America

Ravindra Kasturi