Constructing a table of music similarity vectors from a music similarity graph
First Claim
1. A system for generating a set of coordinate vectors from a sparse graph of media object similarities, comprising using a computing device for:
- receiving a sparse graph of media object similarities;
computing a set of coordinate vectors from the media object similarities of each media object comprising a subset of media objects represented by the sparse graph;
updating the set of coordinate vectors by computing coordinate vectors for each remaining media object represented by the sparse graph which was not included in the subset of media objects; and
storing the set of coordinate vectors to a computer readable storage media.
3 Assignments
0 Petitions
Accused Products
Abstract
A “Music Mapper” automatically constructs a set coordinate vectors for use in inferring similarity between various pieces of music. In particular, given a music similarity graph expressed as links between various artists, albums, songs, etc., the Music Mapper applies a recursive embedding process to embed each of the graphs music entries into a multi-dimensional space. This recursive embedding process also embeds new music items added to the music similarity graph without reembedding existing entries so long a convergent embedding solution is achieved. Given this embedding, coordinate vectors are then computed for each of the embedded musical items. The similarity between any two musical items is then determined as either a function of the distance between the two corresponding vectors. In various embodiments, this similarity is then used in constructing music playlists given one or more random or user selected seed songs or in a statistical music clustering process.
-
Citations
48 Claims
-
1. A system for generating a set of coordinate vectors from a sparse graph of media object similarities, comprising using a computing device for:
-
receiving a sparse graph of media object similarities; computing a set of coordinate vectors from the media object similarities of each media object comprising a subset of media objects represented by the sparse graph; updating the set of coordinate vectors by computing coordinate vectors for each remaining media object represented by the sparse graph which was not included in the subset of media objects; and storing the set of coordinate vectors to a computer readable storage media. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A non-transitory computer-readable medium having computer executable instructions for generating coordinate vectors from a sparse graph of music object similarities, comprising:
-
computing a set of coordinate vectors from music object similarities associated with each music object for each of a set of initial music objects represented by a sparse graph by embedding each initial music object into a multidimensional space as a function of the music object similarities associated with each music object; updating the sparse graph by adding one or more subsequent music objects to the sparse graph; and updating the set of coordinate vectors by computing coordinate vectors for each subsequent music object by holding the coordinate vectors of the initial music objects fixed in the multidimensional space, and iteratively computing a coordinate vector for each subsequent music object as a function of similar initial and subsequent music objects until a convergent embedding solution is achieved. - View Dependent Claims (27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39)
-
-
40. A computer-implemented process for constructing a table of music similarity vectors from a sparse graph of music similarities, comprising:
-
constructing a sparse graph of music similarities representing interrelationships between a plurality of music objects; embedding each music object represented by the sparse graph of music similarities into a multidimensional space by applying multidimensional scaling to the sparse graph, thereby generating an initial set of music similarity vectors from the media object similarities corresponding to each music object; and updating the sparse graph of music similarities by adding one or more subsequent music objects to the sparse graph of music similarities; updating the initial set of music similarity vectors by iteratively generating new music similarity vectors for each subsequent music object while keeping each original music similarity vector in the initial set of music similarity vectors fixed; and storing the updated set of music similarity vectors to a non-transitory computer readable storage media. - View Dependent Claims (41, 42, 43, 44, 45, 46, 47, 48)
-
Specification