University of Pittsburgh

Concept-Enhanced Representation Learning for Linking Heterogeneous Educational Resources

graduate student
Friday, December 1, 2017 - 12:30pm - 1:00pm

In recent years, the education industry has been undergoing an encouraging change. A wide variety of digital contents are openly available to students. Nevertheless, most of these resources are enclosed or weakly connected with each other, making it impossible for learners to find the needed information at the right time. Textbooks, as a well-curated domain knowledge carrier, could form a bond that unites heterogeneous educational resources. This work proposes multiple concept-based contents linking algorithms for connecting online educational resources. Our main focus is on developing an effective representation of domain-specific documents with the help of concepts, by either directly averaging the concept representations or jointly learning with concepts. 

Copyright 2009 | Web site by UMC Web Team