Screen Shot 2014-11-14 at 15.33.19

Recommender systems are commonly used in a variety of applications that include movies, music, news, books and products in general. CrowdRec addresses this demand by providing recommendations that are:

  • Context-aware
  • Resource-combining
  • Interactive
  • Scalable

The key insight of CrowdRec is that, in order to achieve the dense, high-quality, timely information required for such systems, it is necessary to move from collecting information passively from users, to more active techniques. CrowdRec activates the crowd, soliciting input and feedback from the wider community.

Specifically, CrowdRec algorithms create a symbiosis between users and content: they establish reciprocal relationships that both satisfy users’ digital media needs and connect media with users able and willing to contribute the information required to improve access and exchange for the overall community. Additional information is available in our white paper: Activating the Crowd: Exploiting User-Item Reciprocity for Recommendation

The CrowdRec project pursues three objectives:

  • Stream Recommendation: real-time combination of information from collection, context, user interaction and user community to generate social smartfeeds for large-scale social networks;
  • Crowd Engagement: creating symbiosis between users and content that activates users to contribute;
  • Deployment and Validation: developing and testing for release of reference implementations and large-scale user trials.  

For the reference framework containing implementations of algorithms that have been developed within the CrowdRec project,



For a pdf version of the most essential information on what Crowdrec is about,