Services are being developed to aid teachers, librarians, and learners in sharing resources and promoting further access to NSDL resources. The Middleware for Network- and Context-aware Recommendations (MiNC) being developed provides online integrated services for a) understanding the personal activity context through access patterns and analysis of user documents, b) context-aware resource discovery, including search, presentation, and exploration support within the knowledge structure (e.g. Strand Maps) provided by NSDL, and c) peer discovery, peer-network management, and peer driven resource and knowledge sharing and collaborative recommendations.
To support these services, several specific tasks, including representation of the personal activity context, context-aware ranking, filtering, and previews, and integration of context and collaborative filtering are being accomplished. As an integrated service, MiNC operates on existing NSDL digital collection resources as well as on materials from other sites and repositories through available standard interfaces. For accessing NSDL resources, it leverages existing Open Archives Initiative (OAI) based protocols, available web service interfaces based on the Concept Space Interchange Protocol (CSIP), CSIP-Adaptable (CSIP-A) services as well as available NSDL Data Repository APIs. MiNC also integrates with other services, such as IntegraL. The middleware provides API based services to enable future projects and other NSDL contributors to leverage MiNC components, technologies, and information spaces.
The potential exists for MiNC to have a broad educational and social impact. The open interfaces enable future service developers to develop better interfaces targeting general user populations. In addition to the direct educational impact through NSDL, the MiNC system is being integrated into both undergraduate and graduate courses as a project platform. This introduces computer science students to information extraction, information management, recommendation, visualization, and privacy issues as well as to familiarize the students with the important theories related to e-learning.
The principal motivation for this project was to develop content personalization, preview, and collaborative recommendation services that will transform resource sharing and access to online educational content by teachers, librarians, and learners. In particular, we have developed MiNC (a Middleware for Network- and Context-aware Integration and Recommendations on NSDL) which provide sonline integrated services for peer-network driven resource and knowledge sharing and collaborative recommendations. In particular, the outcomes of the MiNC project includes - context-aware resource discovery, including search, presentation, and exploration support within the knowledge structures, - peer discovery, peer-network management, and peer driven resource and knowledge sharing and collaborative recommendations, and - development and deployment of the MiNC peer-networking front-end and back-end in educational and scientific contexts. In particular, MiNC includes a network-aware faceted search system to support user navigation and recommendations within the educational content-/usernetworks: The system involves new algorithms for (a) fast relevant sub-network selection, (b) group-aware incremental PPR computation, (c) in-network target identification, (d) in-network concept-/title- /name-propagation, and (f) facet selection and link-grouping. These included development of analysis techniques to understand temporal evolution of social interactions; including robust uni-variate and multi-variate temporal features that are informed based on the underlying knowledge and peer-graphs. User studies showed that MiNC is able to leverage context to provide contextually relevant recommendations. As part of the project, the project we have developed the core algorithms and the back-end middleware components and deployed and verified these components within the context of scientific and educaitonal peer-networks. We have integrated the various research outcomes into a MiNC middeware, developed a social interface for MiNC middleware, and publically deploy MiNC in educational settings at http://hive.asu.edu/minc/ . In particular, we have deployed the social interface of MiNC in the context of three courses: - CSE408/598, Multimedia Information Systems (Senior/Graduate); 2 instances of the course: + (Fall 12) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=214 + (Fall 13) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=888 - CSE515, Multimedia and Web Databases (Graduate); 3 instances of the course + (Fall 12) http://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=196 + (Fall 13) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=889 + (Fall 14) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=1604 - CSE510, Database Management Systems Internals (Graduate); 2 instances of the course + (Spring 12) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=674 + (Spring 13) https://hive.asu.edu/minc/index.php?option=com_community& view=courses&task=viewcourse&groupid=1360 The project has supported three computer-science PhD students. Three computer-science MS students, both of whom successfully graduated, were also involved in the project. The project team also included two post-doctoral researchers and a PhD student at the U. Torino who collaborated with us within the research context defined by this project. This experience introduced computer science students to research on information extraction, recommendation, and visualization issues related to social media and education. The students also learned how to carry out computational experiments as well as user study based evaluations.
