Dynamic Personalization of Multimedia (MyMedia)
We are drowning in a sea of information overload. Television channels, books and music assault our senses with far too much content. The volume of content on the internet is literally exploding. Not only traditional media but millions of individual users are putting their own content on the web. The massive popularity of YouTube is just one example of this phenomenon. So, in this flood, how do you find content that matters to you? How do you discover multimedia information and entertainment in a way that suits you personally? Isn’t there an easier way? MyMedia recommendations will help solve this “crisis of choice”. Finding what interests you doesn’t have to be an accident. The research project MyMedia is researching solutions that jump beyond traditional recommender systems which are based on a single multimedia source. MyMedia provides recommendations to you that are integrated from many sources. You personalize the system by simply indicating you like a particular video or audio cast and it will find similar content. It will even learn what you like on its own. The more you use it the more it knows your preferences. Personalized recommender technology have the potential to become the central experience for how users access multimedia content. The MyMedia project seeks to advance the state of the art in several areas including creating a software framework for building recommender systems, creating a protocol for plugging in multiple content catalogs, and pluggable recommender algorithms that can be targeted at specific needs. The project will work on new algorithms, new ways to model user preferences, provide the ability to incorporate aspects of social networking to create media centric communities, and research enhancements in the use of metadata for recommender systems. The European Microsoft Innovation Center, BBC Research, BT Research, Microgénesis, Novay and the Universities of Hildesheim and Eindhoven join their expertise in the MyMedia project to pioneer new dynamic personalization software. The resulting system will allow easy integration of multiple content catalogues and recommender algorithms in a single system and provide technology for ranking the content based on personal preferences. The system will learn from user behavior and enable the sharing of recommendation results with friends and family while observing privacy. MyMedia technology will be evaluated on its effectiveness and user friendliness via scientific analysis tools and field trials in several European countries.
Basic Information
Project website: http://www.mymediaproject.org/
Partners:
- European Microsoft Innovation Center (EMIC)
- British Telecommunications plc (BT)
- British Broadcasting Corporation (BBC)
- Novay
- Technical University of Eindhoven (TU/e)
- Microgénesis (MG)
Funding: EU, Small or medium-scale focused research project (STREP)
Contact
Publications
- Zeno Gantner, Lucas Drumond, Christoph Freudenthaler, Lars Schmidt-Thieme (2012):
Personalized Ranking for Non-Uniformly Sampled Items, Journal of Machine Learning Research Workshop and Conference Proceedings .
- Bart P. Knijnenburg, Martijn C. Willemsen, Zeno Gantner, Hakan Soncu, Chris Newell (2012):
Explaining the user experience of recommender systems, User Modeling and User-Adapted Interaction (UMUAI) . - Zeno Gantner, Lucas Drumond, Christoph Freudenthaler, Lars Schmidt-Thieme (2011):
Bayesian Personalized Ranking for Non-Uniformly Sampled Items, in KDD Cup Workshop 2011, San Diego, USA. - Zeno Gantner, Steffen Rendle, Christoph Freudenthaler, Lars Schmidt-Thieme (2011):
MyMediaLite: A Free Recommender System Library, in 5th ACM International Conference on Recommender Systems (RecSys 2011), Chicago, USA. - Christian Wartena, Wout Slakhorst, Martin Wibbels, Zeno Gantner, Christoph Freudenthaler, Chris Newell, Lars Schmidt-Thieme (2011):
Keyword-Based TV Program Recommendation, in 9th Workshop on Intelligent Techniques for Web Personalization and Recommender Systems (ITWP'11), Barcelona, Spain. - Zeno Gantner, Lucas Drumond, Christoph Freudenthaler, Steffen Rendle, Lars Schmidt-Thieme (2010):
Learning Attribute-to-Feature Mappings for Cold-Start Recommendations, in Proceedings of the 10th IEEE International Conference on Data Mining (ICDM 2010), Sydney, Australia. - Zeno Gantner, Steffen Rendle, Lars Schmidt-Thieme (2010):
Factorization Models for Context-/Time-Aware Movie Recommendations, in Challenge on Context-aware Movie Recommendation (CAMRa2010), ACM, Barcelona, Spain. Winner of 'Weekly Recommendation' and 'Live Evaluation' tracks. - Christine Preisach, Leandro Balby Marinho, Lars Schmidt-Thieme (2010):
Semi-Supervised Tag Recommendation - Using Untagged Resources to Mitigate Coldstart Problems, in PAKDD 2010: Proceedings of the 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining. - Steffen Rendle, Christoph Freudenthaler, Lars Schmidt-Thieme (2010):
Factorizing Personalized Markov Chains for Next-Basket Recommendation, in Proceedings of the 19th International World Wide Web Conference (WWW 2010), ACM. Best Paper Award. - Steffen Rendle, Lars Schmidt-Thieme (2010):
Pairwise Interaction Tensor Factorization for Personalized Tag Recommendation, in Proceedings of the Third ACM International Conference on Web Search and Data Mining (WSDM 2010), ACM. Best Student Paper Award. - Zeno Gantner, Christoph Freudenthaler, Steffen Rendle, Lars Schmidt-Thieme (2009):
Optimal Ranking for Video Recommendation, in User Centric Media: First International Conference, UCMedia 2009, Revised Selected Papers, Springer. - Steffen Rendle, Lars Schmidt-Thieme (2009):
Factor Models for Tag Recommendation in BibSonomy, in ECML/PKDD Discovery Challenge 2009 at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML/PKDD-DC 2009). Best Discovery Challenge Award. - Leandro Balby Marinho, Christine Preisach, Lars Schmidt-Thieme (2009):
Relational Classification for Personalized Tag Recommendation, in ECML/PKDD Discovery Challenge 2009 at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML/PKDD-DC 2009). 2nd place in graph-based tag recommendation task. - Paul Marrow, Rich Hanbidge, Steffen Rendle, Christian Wartena, Christoph Freudenthaler (2009):
MyMedia: Producing an Extensible Framework for Recommendation, in Networked Electronic Media Summit 2009. - Zeno Gantner, Lars Schmidt-Thieme (2009):
Automatic Content-based Categorization of Wikipedia Articles, in The People's Web Meets NLP: Collaboratively Constructed Semantic Resources. Workshop at Joint Conference of the 47th Annual Meeting of the Association for Computational Linguistics and the 4th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Processing (ACL 2009). - Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, Lars Schmidt-Thieme (2009):
BPR: Bayesian Personalized Ranking from Implicit Feedback, in Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI 2009). - Steffen Rendle, Leandro Balby Marinho, Alexandros Nanopoulos, Lars Schmidt-Thieme (2009):
Learning Optimal Ranking with Tensor Factorization for Tag Recommendation, in Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining (KDD 2009), Paris, ACM. - Nanopoulos A., Rafailidis D., Ruxanda M., Manolopoulos Y. (2009):
Music Search Engines: Specifications and Challenges, Information Processing & Management . - Ruxanda M., Chua B. Y., Nanopoulos A., Jensen C. (2009):
Emotion-based Music Retrieval on a Well-reduced Audio Feature Space, in Proceedings IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2009), pp. . - Steffen Rendle, Lars Schmidt-Thieme (2008):
Online-Updating Regularized Kernel Matrix Factorization Models for Large-Scale Recommender Systems, in Proceedings of the 2008 ACM Conference on Recommender Systems (RecSys 2008), ACM.