Spotify Ab

The Echo Nest Corporation

Massachusetts Institute of Technology

Nec Laboratories America, Inc.

Brian Whitman

Hui Cao

Daniel Ellis

David P Rodger

Paul B Lamere

Kurt Jacobson

Stephen R Lawrence

Daniel Stowell

Tristan Jehan

Gary William Flake

Tyler Williams

Athena Y Koumis

Jason H Steinbach

Andrew Nesbit

Barry Vercoe

Methods, systems and computer program products are provided for cross-media recommendation by store a plurality of taste profiles corresponding to a first domain and a plurality of media item vectors corresponding to a second domain. An evaluation taste profile in the first domain is applied to a plurality of models that have been generated based on relationship among the plurality of taste profiles and the plurality of media item vectors, and obtain a plurality of resulting codes corresponding to at least one of the plurality of media item vectors in the second domain.

Cross media recommendation

Methods, systems and computer program products are provided for summarizing user activity associated with media content by accessing a taste profile containing a representation of media content activity corresponding to at least one of a plurality of items, generating at least one statistic corresponding to the media content activity, and generating a taste profile attribute by using the at least one statistic.

Taste profile attributes

A catalog record is bridged to information stored in at least one inverted index by receiving an application user interface call associated with a predetermined filter request including a record identifier identifying a record in a relational database. A bitset is generated based on item identifiers in the record. The bitset is applied to at least one inverted index to obtain metadata associated with the item identifiers.

Real time mapping of user models to an inverted data index for retrieval, filtering and recommendation

Methods, systems and computer program products for clustering pages into headline clusters are provided by collecting web data, identifying pages from the web data, tokenizing unique words in each page, recognizing unique entities in each page, detecting media links in each page, and constructing a plurality of vector representations of each page. A first dimension of each vector representation includes the unique words tokenized in each page, a second dimension of each vector representation includes the unique entities recognized in each page, and a third dimension of each vector representation includes the media links detected in each page. The vector representations are, in turn, clustered.

Automatic artist and content breakout prediction

Methods, systems and computer program products are provided for providing content recommendation by obtaining metadata associated with a media object, extracting from the metadata a plurality of terms associated with the media object, and mapping at least a portion of the plurality of terms to buckets. A query vector having attributes corresponding to the buckets is used to perform a query on a database storing media object documents having attributes corresponding to the buckets.

Remote device activity and source metadata processor

A data processing method, program, and apparatus for identifying a document within a block of text. A block of text is tokenized into a plurality of text tokens according to at least one rule parser. Each of the plurality of text tokens is sequentially compared to a plurality of document tokens to determine if the text token matches one of the plurality of document tokens. The plurality of document tokens correspond to a plurality of documents which have been tokenized according to the one or more rule parsers. Each matched text token is filtered according to predetermined filtering criteria to generate one or more candidate text tokens. It is then determined whether sequence of candidate text tokens that occur in sequential order within the block of text match sequence of document tokens. If so, then it is determined that the document has been identified within the block of text. The document can correspond to an artist, a song names, and misspellings and aliases thereof.

Named entity extraction from a block of text

Methods, systems and computer program products are provided for predicting data. A name or title is obtained from a taste profile. There is an index into a data set based on the name or title, and a set of terms and corresponding term weights associated with the name or title are retrieved. A sparse vector is constructed based on the set of terms and term weights. The sparse vector is input to a training model including target data. The target data includes a subset of test data which has a correspondence to a predetermined target metric of data. A respective binary value and confidence level is output for each term, corresponding to an association between the term and the target metric.

Demographic and media preference prediction using media content data analysis

Systems, methods, and computer program products for providing contextually-aware video recommendation

Methods, computing devices, and machine readable storage media for generating a fingerprint of a music sample. The music sample may be filtered into a plurality of frequency bands. Onsets in each of the frequency bands may be independently detected. Inter-onset intervals between pairs of onsets within the same frequency band may be determined. At least one code associated with each onset may be generated, each code comprising a frequency band identifier identifying a frequency band in which the associated onset occurred and one or more inter-onset intervals. Each code may be associated with a timestamp indicating when the associated onset occurred within the music sample. All generated codes and the associated timestamps may be combined to form the fingerprint.

Musical fingerprinting based on onset intervals

A method for fingerprinting an unknown music sample is disclosed. A plurality of known tracks may be segmented into reference samples. A reference fingerprint including a plurality of codes may be generated for each reference sample. An inverted index including, for each possible code value, a list of reference samples having reference fingerprints that contain the respective code value may be generated. An unknown fingerprint including a plurality of codes may be generated from the unknown music sample. A code match histogram may list candidate reference samples and associated scores, each score indicating a number of codes from the unknown fingerprint that match codes in the reference fingerprint. Time difference histograms may be generated for two or more reference samples having the highest scores. A determination may be made whether or not a single reference sample matches the unknown music sample based on a comparison of the time difference histograms.

Growing community of inventors

Cambridge, MA, United States of America

Brian Whitman