System and Method for Natural Language Music Search

• US 20190012381A1
• Filed: 01/26/2018
• Published: 01/10/2019
• Est. Priority Date: 07/06/2017
• Status: Active Grant

• Current Assignee: Sync Floor, Inc.
• Original Assignee: Sync Floor, Inc.
• Inventors: Debique, Kirt
• Granted Patent: US 10,839,008 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/61 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/638 Presentation of query results
• CPC Class Codes: G06F 16/683 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/686 using information manually generated, e.g. tags, keywords, comments, title or artist information, time, location or usage information, user ratings
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

Various methods and systems are provided for implementing a search engine that generates search results for a natural language description of music. A music description and categorization schema is defined to provide a common terminology and taxonomy. Music catalogs are ingested to generate a searchable catalog index. A natural language description of music is analyzed using various natural language processing techniques to generate corresponding musical features of the schema to be searched. A query string is generated from the musical features using a query term profile including one or more query term generators. A ranking score and other search result statistics are generated. Various visualizations can be provided, including a visualization of the translated search and a visualization of search results. In some embodiments, word clouds are provided to allow a user to filter search results by a selected matched feature. As such, users can efficiently review search results.

Claims

1. A computerized method for generating search results from a natural language description of music, the method comprising:

generating an annotated input comprising spans of the description of music and corresponding valence categorizations selected from a set of valence levels;

translating the annotated input to generate corresponding musical features of a music description and categorization schema;

generating a feature map associating the musical features with corresponding ones of the valence categorizations;

generating a query string using the feature map and a query term profile;

executing a search using the generated query string on a catalog index to generate search results and corresponding ranking scores, wherein the catalog index is based on the music description and categorization schema; and

providing the search results using the corresponding ranking scores.

2. The method of claim 1, wherein the generating of the annotated input additionally comprises identifying a musical reference from a reference span of the spans using one or more reference search patterns.

3. The method of claim 1, wherein the translating of the annotated input comprises identifying mentions of musical characteristics from the spans and searching a dictionary for the mentions, wherein the dictionary translates natural language to the musical features.

4. The method of claim 1, additionally comprising applying fuzzing to the generated musical features.

5. The method of claim 1:

wherein the schema comprises categories within which the musical features are distributed;

wherein the query term profile comprises a category query term generator for a first of the categories; and

wherein the generating of the query string additionally comprises;

generating a category query term for ones of the musical features in the category using the category query term generator; and

generating the query string using the category query term.

6. The method of claim 1, additionally comprising normalizing the ranking scores of the search results.

7. The method of claim 1, wherein providing the search results includes presenting a word cloud of matched musical features.

8. One or more computer storage media storing computer-useable instructions that, when used by one or more computing devices, cause the one or more computing devices to perform operations for generating search results from a natural language description of music, the operations comprising:

generating an annotated input comprising spans of the description of music;

translating the annotated input to generate corresponding musical features of a music description and categorization schema;

generating a query string, using the generated musical features as keywords and using a query term profile configured to generate query terms of the query string;

executing a search using the generated query string on a catalog index to generate search results and corresponding ranking scores, wherein the catalog index is based on the music description and categorization schema; and

causing a presentation of the search results using the corresponding ranking scores.

9. The media of claim 8, wherein the generating of the annotated input additionally comprises identifying a musical reference from a reference span of the spans using one or more reference search patterns.

10. The media of claim 8, wherein the translating of the annotated input comprises identifying mentions of musical characteristics from the spans and searching a dictionary for the mentions, wherein the dictionary translates natural language to the musical features.

11. The media of claim 8, wherein the operations additionally comprise applying fuzzing to the generated musical features.

12. The media of claim 8,

wherein the schema comprises categories within which the musical features are distributed;

wherein the query term profile comprises a category query term generator for a first of the categories; and

wherein the generating of the query string additionally comprises;

generating a category query term for ones of the musical features in the category using the category query term generator; and

generating the query string using the category query term.

13. The media of claim 8, wherein the operations additionally comprise normalizing the ranking scores of the search results.

14. The media of claim 8, wherein the presentation of the search results includes a word cloud of matched musical features.

15. A system for generating search results from a natural language description of music, the system comprising:

one or more hardware processors and memory configured to provide computer program instructions to the one or more hardware processors;

a translation component configured to utilize the one or more hardware processors to translate the annotated input to generate corresponding musical features of a music description and categorization schema; and

a matchmaking component configured to utilize the one or more hardware processors to;

generate a query string using the generated musical features and a query term profile; and

execute a search using the generated query string on a catalog index to generate search results and corresponding ranking scores, wherein the catalog index is based on the music description and categorization schema; and

a visualization component configured to utilize the one or more hardware processors to cause a presentation of the search results using the corresponding ranking scores.

16. The system of claim 15, wherein the translation component is additionally configured to identify a musical reference from a reference span of the spans using one or more reference search patterns.

17. The system of claim 15, wherein the translation component is additionally configured to identify mentions of musical characteristics from the spans and search a dictionary for mentions, wherein the dictionary translates natural language to the musical features

18. The system of claim 15, wherein the translation component is additionally configured to apply fuzzing to the generated musical features.

19. The system of claim 15:

wherein the schema comprises categories within which the musical features are distributed;

wherein the query term profile comprises a category query term generator for a first of the categories; and

wherein the generating of the query string additionally comprises;

generating a category query term for ones of the musical features in the category using the category query term generator; and

generating the query string using the category query term.

20. The system of claim 15, wherein the presentation of the search results includes a word cloud of matched musical features.