Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology

Music fans across the Web are now using music streaming services such as Spotify and YouTube for more than downloading sites according to a report in the July 3, 2014 edition of The New York Times entitled Downloads in Decline as Streamed Music Soars. Moreover, streaming is continuing to gain momentum for a number of reasons, not the least of which is that it is less expensive to access. Correspondingly, downloading sales are in decline. This article contains all of the details about this very significant shift in the online music marketplace, including the benefits to consumers and the concerns of musical artists.

A related follow up report entitled How Spotify is Working on Deep Learning to Improve Playlists was posted on Gigaom.com on August 5, 2014 that helps to explain how Spotify is trying to maintain its technical advantage. As reported here, Sander Dieleman, and intern at the company, has developed a method based upon deep learning (a branch of artificial intelligence), that parses large data sets in a new manner aimed at getting newer and lesser known recommendations into Spotify’s users’ playlist recommendations. Underlying this data processing is an examination of the acoustical properties of the user base’s song preferences.

The science behind the recommendation engines used so successfully by Spotify as well as Netflix and Amazon has come along many light years in it sophistication and accuracy since its earliest incarnations on the Web. For a comparative historical perspective, I also, well, recommend checking out an article from the December 1997 issue of WIRED magazine entitled Pattie about the work of Dr. Pattie Maes, a professor at MIT who founded (and later sold to Microsoft), a company called Firefly. Its technology was was one of the first efforts used to create software “agents” to scour the web for user-defined preferences for prices and products.

December 19, 2014 Update: 

Presenting an even stronger case that you-ain’t-seen-nothing-yet in this field was an engaging analysis of some still largely unseen developments in deep learning posted on December 15, 2014, on Gigaom.com entitled What We Read About Deep Learning is Just the Tip of the Iceberg by Derrick Harris. These include experimental systems being tested by the likes of Google, Facebook and Microsoft. As well, there were a series of intriguing presentations and demos at the recent Neural Information Processing Systems conference held in Montreal. As detailed here with a wealth of supporting links, many of these advanced systems and methods are expected to gain more press and publicity in 2015.

 

6 thoughts on “Spotify Enhances Playlist Recommendations Processing with “Deep Learning” Technology

  1. Pingback: Music Visualizations and Visualizatons About Music |

  2. Pingback: Startup is Visualizing and Interpreting Massive Quantities of Daily Online News Content |

  3. Pingback: Is Big Data is Calling and Calculating the Tune in Today’s Global Music Market? |

  4. Pingback: Updates on Recent Posts Re: Music’s Big Data, Deep Learning, VR Movies, Regular Movies’ Effects on Our Brains, Storytelling and, of Course, Zombies |

  5. Pingback: Three New Perspectives on Whether Artificial Intelligence Threatens or Benefits the World |

  6. Pingback: Rock It Science: New Study Equates Musical Tastes to Personality Traits |

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s