Supercomputing and Computation


Personalized Content Recommendations

Problem Statement

  • How to detect user interests and automatically recommend interesting contents in a personalized way.

Technical Approach

  • Detecting user interests through attention time, i.e. time spent by a user on reading a certain webpage.
  • Collaboratively mining semantic contents of user reading materials along with one’s implicit feedbacks.
  • Advanced data fusion algorithms for user interest inference.
  • Dynamic content recommendations according to inferred user interest profile.

Advantage over the State-of-the-Art

  • Leverage an ontology based approach for noise tolerant user interest inference.
  • Can autonomously recommend interesting contents to end users without explicit user participation.
  • Capable of detecting dynamic user interest shift fully automatically and adjust algorithm behaviors accordingly.

Prototype website:


We're always happy to get feedback from our users. Please use the Comments form to send us your comments, questions, and observations.