Material Detail

Natural language processing supported transdisciplinary crowdsourcing

Natural language processing supported transdisciplinary crowdsourcing

This video was recorded at Video Journal of Machine Learning Abstracts - Volume 4. We outline a Web-based architecture that serves the construction of transdisciplinary knowledge and is supported by machine intelligence. At the heart of the architecture is the interaction between a repository of concepts and the machine intelligence. The repository is an ontology integrated with a Wiki; each concept in the ontology is grounded in the natural language text of the Wiki. The machine intelligence exploits structured sparse coding and helps users interact with the concept repository. Most importantly, it helps practitioners of different fields understand each other by explaining unknown terminology in the Wiki, and it facilitates creating and maintaining content. These two components evolve together: as the concept repository grows, the intelligence performs better. Furthermore, extending the concept repository semi-automatically becomes easier. To support wide contribution and ensure the high quality of the accumulated knowledge, content is divided into two types: drafts and articles. Anyone who registers can create and edit drafts, but only drafts that pass a voting procedure and are approved by an Editorial Board can become articles. Articles are very similar to scientific papers: they undergo peer-review and contain verified knowledge. Two communities have already started using these portals in the domains of robotic surgery and education.


  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material


Disciplines with similar materials as Natural language processing supported transdisciplinary crowdsourcing


Log in to participate in the discussions or sign up if you are not already a MERLOT member.