Presentation is loading. Please wait.

Presentation is loading. Please wait.

ITunes Genius Presented By: Dibyendu Talukder (MT13063) Prerna Juneja (MT13099)

Similar presentations


Presentation on theme: "ITunes Genius Presented By: Dibyendu Talukder (MT13063) Prerna Juneja (MT13099)"— Presentation transcript:

1 iTunes Genius Presented By: Dibyendu Talukder (MT13063) Prerna Juneja (MT13099)

2 Overview  Introduction  How it Looks  Idea  Working  Limitations  References 10/29/20142

3 Introduction iTunes Genius is a feature of iTunes that provides two main features: i) Creating playlist from your existing library. ii) Suggesting new music at the iTunes Store based on the music you like in your iTunes library. 10/29/20143 iTunes Genius Logo

4 How It Looks!! 410/29/2014

5 Idea  iTunes Genius functionality was added to iTunes with version 8.0, release on Sept. 9, 2008.  Genius Mixes feature has been added with the version iTunes 9.  Genius Shuffle feature has been added with iTunes 11.1. 10/29/20145

6 Genius Playlists:  Automatically creates playlists of songs you already own that will sound great together using collaborative filtering techniques.  The starting point for the Genius service: Gathering & sending information about your iTunes library to Apple’s servers. 10/29/20146 How It Works

7 10/29/20147

8 8 First stage of the upload is likely using their fuzzy text matching to link each of our text to the info they have from the iTunes store. Two of them have "Dream On" by Aerosmith & Boplicity by Rolling Stones songs in their collections. But that doesn't make them so "similar" at this point.

9  Apple applies a layer of filters to trim down the lists based on:-  Genre  BPM (Beats Per Minute)  At this point Apple has a whole huge list of songs in the same genre.  Prioritize the list by sales rank numbers or overall popularity based on the number of times they've seen the song in other peoples' collections => Master List 10/29/20149

10  Problem: List likely to have a bunch of songs in the list that I don't have in my collection (and therefore can't show up in a playlist).  Apple needs to reconcile their Master List with the info from my collection, and then sends that back.  Generates (at least) 100-song static playlists for each eligible song, since one can bump the number of songs up from 25 to 100 without making another call up to the servers 10/29/201410

11 10/29/201411 Ranking of the songs tries to ensure that the best matches come back most often near the top.

12 Discover New Music: The library of tracks is compared to all the other Genius users’ libraries of tracks. Apple then runs a set of recommendation algorithms to suggest songs for a user to add to his queue. Statistics are computed globally at regular intervals & stored in a cache Reason: Data on the similarity of any two songs changes slowly– only reason it changes at all is because of the changing tastes of the listening public, and the introduction of new tracks and artists. 1210/29/2014

13  Recommendation methods uses information retrieval algorithms.  TF-IDF method is used to assign weightage to factors to determine their importance.  Way to compare how often a particular factor occurs in a single library to how often that factor occurs in a larger body such as the sum of all iTunes libraries stored by the Genius servers. 10/29/201413

14  tf-idf weights are sorted and represented in a vector space model as vectors.  D1, D2=> tf-idf weights of songs represented as a single vector.  query (q)=>“which of these two songs is most like the one for which I’ve just clicked the ‘genius’ button.”  Whichever one is closer in angle to your query vector is more similar. 10/29/2014 14

15 Limitations  Large disk space requirements for a media player.  Library Genius.itdb which is roughly 5 times the size of iTunes library.  Recommendations are not always perfect. You might not like some artists that other users with a slightly similar library enjoy. 1510/29/2014

16 References 1.datawhat.blogspot.com.au/2008/09/real-genius.html 2.www.technologyreview.com/view/419198/how-itunes-genius- really-works/ 3.http://ipod.about.com/od/itunes/g/itunes_genius.htmhttp://ipod.about.com/od/itunes/g/itunes_genius.htm 4.http://en.wikipedia.org/wiki/Tf%E2%80%93idfhttp://en.wikipedia.org/wiki/Tf%E2%80%93idf 5.http://www.pcmag.com/article2/0,2817,2368920,00.asp 1610/29/2014

17 THANK YOU!!!


Download ppt "ITunes Genius Presented By: Dibyendu Talukder (MT13063) Prerna Juneja (MT13099)"

Similar presentations


Ads by Google