1st Workshop on Recommender Systems for TV and Online Video – RecSysTV 2014

RecSysTV Program – 10th October 2014, Foster City, Silicon Valley, USA

Morning Session

  • 10:30-11:00 Morning Break

Mid-Morning Session

  • 11:00-11:15 A Graph-based Collaborative and Context-aware Recommendation System for TV Programs, E. Şamdan et al.
  • 11:15-11:30 Time-based TV Programs Prediction, R. Turrin et al.
  • 11:30-11:45 Augmented Matrix Factorization with Explicit Labels for Recommender Systems, J. Zhou et al.

Lunch Demo/Poster Sessions

  • 11:45-12:15 Comcast, Boxfish, GraphLab Presentations
  • 12:15-14:00 Lunch with Posters & Demonstrations from Attendees

Posters

  • R. Turrin – Time-based TV Programs Prediction
  • H. Corona – A Mood-based Genre Classification of Television Content
  • J. Kuchař – Bag-of-Entities Text Representation for Client-side TV Recommender Systems
  • K. Yamamoto – Content-based Viewer Estimation Using Image Features for Recommendation of Video Clips
  • A. Taşci – A Graph-based Collaborative and Context-aware Recommendation System for TV Programs
  • M. Sharma – Feature-based Bilinear Similarity Model for Top-n Recommendation of New Items

Demos

  • J. Neumann – Comcast
  • J. Hannon – Boxfish
  • D. Bickson – GraphLab
  • B. Kitts – Adap.tv
  • D. Zibriczky – ImpressTV
  • P. Auteri – Contentwise
  • J. Delgado – OnCue TV

Afternoon Session

  • 14:00-14:45 (Invited Talk) Building Large-Scale Recommender Systems for TV, J. Hu (Samsung)
  • 14:45-15:00 Item-Based Collaborative Filtering Using the Big Five Personality Traits, H. Alharthi et al.
  • 15:00-15:15 The Cold Start: Minimal User’s Rating Activity Estimation, O. Kaššák et al.
  • 15:15-15:30 A Mood-based Genre Classification of Television Content, H. Corona Pampín et al.
  • 15:30-16:00 Afternoon Break

Mid-Afternoon Session

  • 16:00-16:15 Bag-of-Entities text representation for client-side TV recommender systems, J. Kuchař et al.
  • 16:15-16:30 Content-based Viewer Estimation Using Image Features for Recommendation of Video Clips, K. Yamamoto et al.
  • 16:30-16:45 Pilot User Selection Method for Joint-Prediction of Ratings and Popularity on Cold-Start Item, Z. Miao et al.

  • 16:45-17:00 Wrap Up and Awards

 

Organizing committee

Danny Bickson, Graphlab Inc., Seattle, WA

John Hannon, Boxfish, Palo Alto, CA

Jan Neumann, Comcast Labs, Washington, DC

Hassan Sayyadi, Comcast Labs, Washington, DC

Program committee:

Justin Basilico, Netflix

Balázs Hidasi, GravityR&D

Craig Carmichael, Rovi

Emanuele Coviello, Keevio

Pádraig Cunningham, Insight Centre for Data Analytics

Joaquin Delgado, OnCue TV (Verizon)

Diana Hu, OnCue TV (Verizon)

Ben Jordan, Univ. of Minnesota

Brendan Kitts, Adap.tv (AOL)

Noam Koenigstein, Microsoft

Gert Lanckriet, UC San Diego

Hayim Makabee, Yahoo! Labs

Tom Rampley, Dish Network

Royi Ronen, Microsoft

Barry Smyth, Insight Centre for Data Analytics

Domonkos Tikk, GravityR&D

Ari Tuchman, quantiFind

Udi Weinsberg, Technicolor Labs

Esti Widder, Viaccess-Orca

Ho-Hsiang Wu, Rd.io

Dávid Zibriczky, ImpressTV

Jiayu Zhou, Samsung Research

Sponsored By:

Comcast Labs is looking for a Computer Vision Lead Researcher

Are you interested in solving computer vision problems that involve massive video sets and thousand channels of live TV? Do you have experience in building large-scale computer vision and machine learning system?

Do you want to conduct industry-leading research in content analysis technologies and multimedia information processing, and help millions of households discover and enjoy video and music content on their TV, PC, Phone, and Mobile devices?

Comcast Labs in Washington DC is currently looking to fill a lead researcher position that requires computer vision and video processing research and prototype development applied to premium video, live TV and home security.

We are an innovative research group within Comcast’s Metadata Products and Search Services unit that does groundbreaking research to develop Search & Discovery technologies for Video and TV that support Comcast’s approximately 22 Million subscribers. Comcast is the largest provider of TV and Broadband Services in North America, the largest provider of TV search and discovery applications in North America, & the 6th largest provider of search on the web.
The ideal candidate will have experience working in an industrial, government, or academic lab setting on computer vision and video processing projects and applicants with PhD are preferred. Applicants should have good programming and software development skills, and be comfortable working in an interdisciplinary, team-oriented, applied research environment.

The position will be located in our Washington, DC office. We feature an informal, open atmosphere, and a location in downtown Washington convenient to several public transit lines and many of the city’s museums, monuments, and other attractions. The salary is competitive and commensurate with experience.
To apply, please send a CV or resume, along with a brief statement explaining why you are interested in the position, to Jan_Neumann(at)cable.comcast.com.

Core Responsibilities:
– Develops specifications and technical requirements of custom designs for future products and applications.
– Works with other Technical Leads, Product Managers and Business Partners to lead development of prototypes that demonstrate Research work and help with Product Discovery.
– Works with various team members both within and outside Research. Ensures timely progress of work. Conducts Research in an incremental manner thereby enabling faster transfer of technology to the Engineering teams. Able to evaluate prototype systems, help write technical papers, and help with technology transfer.
– Keeps track of developments in field both in academia and in industry. Attending relevant conferences and publishing research results is encouraged.
– Other duties and responsibilities as assigned. Regular, consistent and punctual attendance. Must be able to work nights and weekends, variable schedule(s) as necessary.

Education Level: Master’s Degree or Equivalent; Ph.D. preferred.
Field of Study: Computer Science with experience in one or more of the following areas Information Retrieval, Machine Learning, Computer Vision, Video Processing.
Years of Experience: Generally requires 7-10 years related experience after Bachelors, 5-8 years after Master’s, or 2-3 years after Ph.D.

Compliance: Comcast is an EEO/AA/Drug Free Workplace.
Disclaimer: The above information has been designed to indicate the general nature and level of work performed by employees in this role. It is not designed to contain or be interpreted as a comprehensive inventory of all duties, responsibilities and qualifications

Comcast Labs DC is looking for 4-5 Graduate Student Summer Interns

Are you interested in Large-scale Machine Learning? Multimedia Video Processing? Working with truly Big Data? Do you currently do research in at least one or more of the following areas – Large-scale Machine Learning, Recommendation Systems, Social Media, Computer Vision, Information Retrieval or NLP? Do you want to conduct industry-leading research in content discovery technologies and multimedia information processing, and help millions of households discover video and music content on their TV, PC, Phone, and Mobile devices?

Comcast Labs in Washington DC is currently looking to fill 4-5 graduate student intern positions for this summer (minimum of 12 weeks, May through September). Projects can focus on using Statistical Modeling of TV Viewing Behaviors, Personalized Recommendations for TV, and Search, Annotation, and/or Segmentation of premium video.

We are an innovative research group within Comcast’s Metadata Processing and Search Services unit that does groundbreaking research to develop the video and TV search & discovery technologies that support Comcast’s approximately 22 Million subscribers.

Comcast is the largest provider of TV and Broadband Services in North America, the largest provider of TV search and discovery applications in North America, & the 6th largest provider of search on the web.

The ideal applicant will be currently enrolled in a university PhD program and has 2+ years of research experience in one of the relevant areas. Applicants should also have good programming and software development skills, and be comfortable working in an interdisciplinary, team-oriented, applied research environment.

Internships will be on-site in our Washington, DC office. We feature an informal, open atmosphere, and a location in downtown Washington convenient to several public transit lines and many of the city’s museums, monuments, and other attractions. The salary is competitive and commensurate with experience.

To apply, please send a CV or resume, along with a brief statement explaining why you are interested in the position, to Jan_Neumann(at)cable.comcast.com. While there is no fixed deadline for applying, we anticipate filling these positions by the end of February.

 

 

Talk on Popularity Prediction of TV Shows at NCTA Technical Forum

I presented a talk at the yearly meeting of the National Cable and Television Association (NCTA) on how to combine TV usage statistics with social signals such as Twitter for improved popularity prediction of TV Shows. Here is the link to the paper. (Also, see here for an interesting take on our use of random forests :-) http://www.multichannel.com/blog/translation-please/best-ncta-s-tech-papers-part-2/373214)

2010 Summer Workshop of the Center for Language and Speech Processing

During the summer of 2010 I led a research team at the 2010 Summer Workshop of the Center for Language and Speech Processing at Johns Hopkins University. For 6 weeks we worked on combining text and video analysis to identify and localize complex actions in broadcast videos. You can find more detailed information about our project and our results here.