internet new web2.0 trends-recommendation systems lecture for vtlv

22
MEDIA RECOMMENDATION SYSTEMS Yossi Cohen CTO DSP-IP

Upload: smo

Post on 13-Jan-2015

4.195 views

Category:

Technology


0 download

DESCRIPTION

Internet New Web2.0 Trends-Recommendation Systems, Lecture For Vtlv conference Presentation the new internet media recommendation systems principles collaborative filtering and content based filtering. examples of recommendation services for music and video like: pandora, lastfm, video: flixter...etc more at www.dsp-ip.com

TRANSCRIPT

Page 1: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MEDIA RECOMMENDATION SYSTEMSYossi CohenCTODSP-IP

Page 2: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

AGENDA

• Why Do we need it?• What are recommendation systems?

– Content based– Collaborative– Hybrid

• Examples– Music – Video

• Summary & Trends

Page 3: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

WHY RECOMMENDATION ENGINE?

Page 4: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

RECOMM. SYSTEM DEFINITION

• Recommender systems are a specific type of information filtering (IF) technique that attempt to present to the user information items (movies, music, books, news, web pages) the user is interested in.

• To do this the user's profile is compared to some reference characteristics. These characteristics may be from the information item (the content-based approach) or the user's social environment (the collaborative filtering approach).

Page 5: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

TAXONOMY

Page 6: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

COLLABORATIVE FILTERING

• Collaborative filtering (CF) is the method of making automatic predictions (filtering) about the interests of a user by collecting taste information from many users (collaborating). The underlying assumption of CF approach is that those who agreed in the past tend to agree again in the future. For example, a collaborative filtering or recommendation system for music tastes could make predictions about which music a user should like given a partial list of that user's tastes (likes or dislikes).

Source: TrustedOpinion

Page 7: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MUSIC RECOMMENDATION ENGINE

Page 8: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MUSIC RECOMMENDATION SERVICES

Page 9: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

LAST.FM

Page 10: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MEEMIX

• Try to analyze the user taste in order to provide the best personalized music channel

Page 11: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

VIDEO RECOMMENDATION ENGINE

Page 12: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MOVIELENS

• University research project• Web 0.5 GUI• You rate• Get Predictions

Page 13: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

FLIXSTER

• Commercial version of MovieLens with better features & GUI

Collaborative /Peer based

Content based

Page 14: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

SUMMARY

Page 15: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

RECOMMENDATION SYSTEM IN

• Content Creators : – Disney, HBO

• Social Web sites : – FaceBook (MyTV/Video apps), MySpace

• Content Aggregators– Magnify, Dabble, SuTree(Israel), Nebo

• Internet (Content) Recommenders– BFN(Alpha-Israel), Mogad, MyStrends

• Mobile Content Recommenders– JumpTap, Matchmine

Page 16: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

MORE PLACES FOR RECOM. SYSTEMS

•Channel Creation Platforms

•Video Mesh-ups

Page 17: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

TRENDS

ProcessPastPresentFuture

DistributionBroadcastPre defined

channels

UGC / Pre defined channels

My Customized channel

ConsumptionLean Back – open loop

Lean ForwardText search of clips

Lean back

Content typePremium ContentUGC content + Stolen Premium

content

UGC + Legal Premium

content+Meshups

Page 18: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

TRENDS

• Recommendation Systems closes the loop between content creator/distributers and the users

Page 19: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

SUMMARY• Recommendation engine will not stay only as a stand alone system like

Google in general search• Evolve into serving platforms to provide lean-back / personal channel user

experience like: Pandora, MeeMix, LastFM• Immersed into content aggregation and distribution platforms• Work as a portal (Google) or a download application (Last.FM, Veoh,

MatchMine)• Use a lot of Flash, AIR, Apollo

Page 20: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

DSP-IP SERVICES

• Outsourcing, consulting and development services– Video encoding and streaming– Flash Video, VP6– IPTV architecture and services– Recommendation System consulting– Video on DSP platforms (TI DSPs)– Video Advertisement in IPTV, Internet and mobile– Image processing

Page 21: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

DSP-IP DSP-IP CONTACT INFORMATION

www.dsp-ip.com Giborey Israel 20, POB 8323, Netanya, IsraelGiborey Israel 20, POB 8323, Netanya, IsraelOffice Phone: 09-8850956, Fax: 050- 8962910Office Phone: 09-8850956, Fax: 050- 8962910

HR Services: HR Services: Michal PoratMichal [email protected], 054-238368909-8651933, 054-2383689

Technology Management Technology Management Services Services : : Yossi CohenYossi [email protected], 054-5313092 09-8850956, 054-5313092

Page 22: Internet New Web2.0 Trends-Recommendation Systems Lecture For Vtlv

RESOURCES

• Recommendation Systems http://en.wikipedia.org/wiki/Recommendation_system • Collaborative Filtering http://en.wikipedia.org/wiki/Collaborative_filtering • MeeMix www.meemix.com • MovieLens http://movielens.umn.edu/main• Flixster http://www.flixster.com/ • MyStrends www.mystrends.com

http://www.moconews.net/entry/419-recommendation-engine-provider-mystrands-receives-25-million-in-funding/

• Matchmine www.matchtime.com • http://www.killerstartups.com/Video-Music-Photo/matchmine--Media-Discovery-Platform/