Methods and systems for personalizing user experience based on diversity metrics

Patent metadata

Field	Value
Publication number	US-10133545-B2
Application number	US-201715729351-A
Country	US
Kind code	B2
Filing date	Oct 10, 2017
Priority date	Feb 24, 2017
Publication date	Nov 20, 2018
Grant date	Nov 20, 2018

Abstract

A server system of a media-providing service has one or more processors and memory that stores instructions for execution by the one or more processors. The server system accesses a listening history of a user that specifies tracks consumed by the user. The tracks have respective characteristics that describe the tracks. The server system determines for the user, using the listening history, a diversity score based on variation of the characteristics across the tracks consumed by the user. The server system provides personalized content to the user based on the diversity score.

First claim

What is claimed is:

1. A method, comprising,
at a server system of a media-providing service having one or more processors and memory storing instructions for execution by the one or more processors:
accessing a listening history of a user that specifies tracks consumed by the user, wherein:
the listening history includes:
a current time window that includes a current time, and
a previous time window that includes a previous time that is prior to the time span of the current time window; and
the tracks have respective characteristics that describe the tracks;

determining for the user, using the listening history, a diversity score based on variation of the characteristics across the tracks consumed by the user, wherein determining the diversity score includes determining a current aggregate diversity score based on variation of the characteristics for pairs of tracks within the current time window, wherein the pairs of tracks within the current time window include sequential and non-sequential pairs of tracks within the current time window;
comparing the current aggregate diversity score of the current time window with a historical aggregate diversity score of the previous time window, wherein the historical aggregate diversity score is based on variation of the characteristics for pairs of tracks within the previous time window, wherein the pairs of tracks within the previous time window include sequential and non-sequential pairs of tracks within the previous time window;
determining whether the diversity score satisfies a first threshold;
in response to determining that the diversity score satisfies the first threshold, reordering a set of tracks to provide the set of tracks to the user in a selected order, wherein the selected order increases a jump distance between successive tracks in the set in order to increase a degree of diversity between the successive tracks; and
providing, to a client device of the user for display in a user interface displayed by the client device, personalized content based on the current aggregate diversity score and the comparing of the current aggregate diversity score of the current time window with the historical aggregate diversity score of the previous time window, the providing including providing the tracks from the reordered set in the selected order.

2. The method of claim 1 , wherein:
the listening history of the user comprises a plurality of sessions for the user;
the current time window corresponds to a current session for the user;
the previous time window includes one or more previous sessions; and the historical aggregate diversity score is based on variation of the characteristics for pairs of tracks within the one or more previous sessions.

3. The method of claim 1 , wherein:
the listening history further specifies an order in which the user consumed the tracks;
the diversity score includes a sequential diversity score; and
determining the diversity score further comprises determining the sequential diversity score based on variation of the characteristics for sequential pairs of tracks in the listening history.

4. The method of claim 1 , wherein:
the respective characteristics of the tracks are represented by respective collaborative filtering vectors in a latent space; and
determining the diversity score comprises measuring distances between the collaborative filtering vectors in the latent space.

5. The method of claim 1 , wherein the respective characteristics that describe the tracks include values associated with categories selected from the group consisting of genre, artist, popularity, playlist co-occurrence, and emotion.

6. The method of claim 1 , further comprising, at the server system:
determining whether the diversity score satisfies a second threshold;
in response to determining that the diversity score satisfies the second threshold, increasing a degree of diversity associated with the user; and
selecting a track with one or more characteristics that differ from one or more characteristics of a current track by no more than the increased degree of diversity;
wherein providing the personalized content comprises providing the selected track.

7. The method of claim 1 , further comprising, at the server system:
determining whether the diversity score satisfies a second threshold;
in response to determining that the diversity score does not satisfy the second threshold, decreasing a degree of diversity associated with the user; and
selecting a track with one or more characteristics that differ from one or more characteristics of a current track by no more than the decreased degree of diversity;
wherein providing the personalized content comprises providing the selected track.

8. The method of claim 1 , further comprising, at the server system:
in response to determining that the diversity score does not satisfy the first threshold, reordering the set of tracks to decrease the jump distance between successive tracks in the set in order to decrease a degree of diversity between the successive tracks.

9. The method of claim 1 , wherein providing the personalized content comprises determining whether to recommend a track to the user based on the diversity score.

10. The method of claim 1 , further comprising, at the server system, selecting a content source based at least in part on the diversity score, wherein:
the content source is at least one of a station or a pre-established playlist; and
providing the personalized content comprises recommending the selected content source to the user.

11. The method of claim 1 , wherein providing the personalized content comprises recommending personalized content to the user based on the diversity score.

12. The method of claim 1 , wherein providing the personalized content comprises providing a message to the user.

13. The method of claim 12 , further comprising, at the server system:
determining whether the diversity score satisfies a third threshold; and
in response to determining that the diversity score satisfies the third threshold, selecting the message.

14. A server system of a media-providing service, comprising:
one or more processors; and
memory storing one or more programs for execution by the one or more processors, the one or more programs comprising instructions for:
accessing a listening history of a user that specifies tracks consumed by the user, wherein:
the listening history includes:
a current time window that includes a current time, and
a previous time window that includes a previous time that is prior to the time span of the current time window; and
the tracks have respective characteristics that describe the tracks;

determining for the user, using the listening history, a diversity score based on variation of the characteristics across the tracks consumed by the user, wherein determining the diversity score includes determining a current aggregate diversity score based on variation of the characteristics for pairs of tracks within the current time window, wherein the pairs of tracks within the current time window include sequential and non-sequential pairs of tracks within the current time window;
comparing the current aggregate diversity score of the current time window with a historical aggregate diversity score of the previous time window, wherein the historical aggregate diversity score is based on variation of the characteristics for pairs of tracks within the previous time window, wherein the pairs of tracks within the previous time window include sequential and non-sequential pairs of tracks within the previous time window;
determining whether the

Assignees

• Spotify Ab

Inventors

• Gibson Clay
• Shapiro Will
• Gil Santiago
• Anderson Ian
• Mpossi Margreth
• Semerci Oguz
• Wolf Scott

Classifications

• H04L67/306

  User profiles · CPC title

• H04N21/2407

  Monitoring of transmitted content, e.g. distribution time, number of downloads (arrangements for monitoring programmes for broadcast or for distribution combined with broadcast H04H20/14) · CPC title

• H04N21/26258

  for generating a list of items to be played back in a given order, e.g. playlist, or scheduling item distribution according to such list (retrieval of multimedia data based on playlists G06F16/40) · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/635

  Filtering based on additional data, e.g. user or group profiles · CPC title