Multi-domain search on a computing device

• US 10,642,845 B2
• Filed: 09/30/2014
• Issued: 05/05/2020
• Est. Priority Date: 05/30/2014
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Hornkvist, John M., Kapoor, Gaurav
• Primary Examiner(s): Alam, Hosain T
• Assistant Examiner(s): May, Robert F
• Application Number:
  US14/502,881
• Publication Number:
  US 20150347594A1
• Time in Patent Office: 2,044 Days
• Field of Search: 707728
• CPC Class Codes: G06F 16/2457 with adaptation to user needs
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9537 Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

Systems and methods are disclosed for improving search results returned to a user from one or more domains, utilizing query features learned locally on the user'"'"'s device. One or more domains can inform a computing device of one or more features related to a search query upon which to the computing device can apply local learning. A local search system can include a local database, a local search history and feedback history database, and a local learning system to identify features about query terms. The features can be learned from the user'"'"'s interaction with both local search results and remote search results, without sending the user interaction information or other user identification information to a remote search engine. A locally learned feature can be used to extend a query, bias a query term, or filter query results.

Claims

1. A computer-implemented method, comprising:

transmitting, by a computing device of a user, to a remote search engine and to a local database, a search query;

receiving, by the computing device, from the remote search engine and the local database, search results responsive to the search query;

identifying, by a local learning system on the computing device, a feature related to the search query, using user interaction with the search results received from the local database, the remote search engine, or both, and wherein the user interaction with the search results is stored on the computing device of the user and kept private to the computing device of the user and is not shared with the remote search engine, wherein the feature comprises a distinction of multiple meanings of a term of the search query or a term in the search results;

generating, on the computing device, a local predictor for the feature identified by the local learning system;

training the local predictor on the computing device for the feature using information generated on, and private to, the computing device, wherein the trained local predictor indicates a learned user bias towards one of the multiple meanings of the term of the search query or the term of the search results;

in response to the local learning system identifying the feature, the computing device transmitting a feature vector containing the locally identified feature to the remote search engine, wherein the feature vector is distinct from the search query, the feature vector does not contain an indication of bias towards one of the multiple meanings related to the identified feature, and the remote search engine is to generate and train a predictor for the remote search engine for the feature identified by the local learning system; and

transmitting a second query to the remote search engine and local database, the second query having one or more terms related to the identified feature, and receiving second query results from the remote search engine and local database based on the multiple meanings related to the identified feature.

2. The method of claim 1, further comprising biasing a subsequent query to the remote search engine and the local database in accordance with the trained predictor.

3. The method of claim 1, wherein information obtained from the computing device comprises at least one of:

a search query performed on the computing device of information on the computing device or feedback of interaction by a user of the computing device with results returned from a search query performed on the computing device of information stored on the computing device, wherein the feedback of interaction by the user of the computing device is not transmitted to the remote search engine.

4. The method of claim 3, wherein training comprises a statistical analysis of the information obtained from the computing device, wherein statistical analysis comprises one of:

linear regression, Bayes classification, or Naive Bayes classification.

5. The method of claim 1, wherein information generated on the computing device comprises a context of the computing device at the time the search query was issued.

6. The method of claim 5, further comprising:

wherein the context includes an application from which the search query was issued.

7. The method of claim 1, further comprising:

anonymizing, by the computing device of the user, the user interaction history with the remote search engine results, such that a particular user and/or particular computing device of a particular user is not identified with the anonymized user interaction history; and

transmitting the anonymized user interaction history with the remote search engine results to the remote search engine.

22. The method of claim 1, further comprising:

receiving, by the computing device, local search results responsive to the search query;

blending and ranking, by the computing device, the local search results with the remote search engine results; and

displaying the blended search results to a user on the computing device.

8. A non-transitory computer-readable medium programmed with executable instructions that, when executed by a processing system, perform a method, comprising:

transmitting, by a computing device of a user, to a remote search engine and a local database, a search query;

receiving, by the computing device, from the remote search engine and the local database, search results responsive to the search query;

identifying, by a local learning system on the computing device, a feature related to the search query, using user interaction with the search results received from the local database, the remote search engine, or both, and wherein the user interaction with the search results is stored on the computing device of the user and kept private to the computing device of the user and is not shared with the remote search engine, wherein the feature comprises a distinction of multiple meanings of a term of the search query or a term in the search results;

generating, on the computing device, a local predictor for the feature identified by the local learning system;

training the local predictor for the feature on the computing device, using information generated on, and private to, the computing device, wherein the trained local predictor indicates a learned user bias towards one of the multiple meanings of the term of the search query or the term of the search results;

in response to the local learning system identifying the feature, the computing device, transmitting a feature vector containing the locally identified feature to the remote search engine, wherein the feature vector is distinct from the search query, the feature vector does not contain an indication of bias towards one of the multiple meanings related to the identified feature and the remote search engine is to generate and train a predictor for the remote search engine for the feature identified by the local learning system; and

transmitting a second query to the remote search engine and local database, the second query having one or more terms related to the identified feature, and receiving second query results from the remote search engine and local database based on the multiple meanings related to the identified feature.

9. The medium of claim 8, further comprising biasing a subsequent query to the remote search engine and the local database in accordance with the trained predictor.

10. The medium of claim 8, wherein information obtained on the computing device comprises at least one of:

a search query performed on the computing device of information on the computing device or feedback of interaction by a user of the computing device with results returned from a search query performed on the computing device of information stored on the computing device, wherein the feedback of interaction by the user of the computing device is not transmitted to the remote search engine.

11. The medium of claim 10, wherein training comprises a statistical analysis of the information obtained from the computing device, wherein statistical analysis comprises one of:

linear regression, Bayes classification, or Naive Bayes classification.

12. The medium of claim 8, wherein information generated on the computing device comprises a context of the computing device at the time the search query was issued.

13. The medium of claim 12, further comprising:

wherein the context includes an application from which the search query was issued.

14. The medium of claim 9, further comprising:

anonymizing, by the computing device of the user, user interaction history with the remote search engine results, such that a particular user and/or particular computing device of the particular user is not identified with the anonymized user interaction history; and

transmitting the anonymized user interaction history with the remote search engine results to the remote search engine.

15. A system of a user comprising:

a memory; and

a processing system programmed with executable instructions stored in the memory that, when executed by the processing system, perform a method, comprising;

transmitting, by the system, to a remote search engine and a local database, a search query;

receiving, by the system, from the remote search engine and the local database, search results responsive to the search query;

identifying, by a local learning system of the system of the user, a feature related to the query, using user interaction with the search results received from the local database, the remote search engine, or both, and wherein the user interaction with the search results is stored on the system of the user and kept private to the system of the user and is not shared with the remote search engine, wherein the feature comprises a distinction of multiple meanings of a term of the search query or a term in the search results;

generating, on the system, a local predictor for the feature identified by the local learning system;

training the local predictor for the feature on the system of the user, using information generated on, and private to, the system, wherein the trained local predictor indicates a learned user bias towards one of the multiple meanings of the term of the search query or the term of the search results;

in response to local learning system identifying the feature, the system of the user transmitting a feature vector containing the locally identified feature to the remote search engine, wherein the feature vector is distinct from the search query, the feature vector does not contain an indication of bias towards one of the multiple meanings related to the identified feature, and the remote search engine is to generate and train a predictor for the remote search engine for the feature identified by the local learning system; and

transmitting a second query to the remote search engine and local database, the second query having one or more terms related to the identified feature, and receiving second query results from the remote search engine and local database based on the multiple meanings related to the identified feature.

16. The system of claim 15, further comprising biasing a subsequent query to the remote search engine and the local database in accordance with the trained predictor.

17. The system of claim 15, wherein information obtained on the system comprises at least one of:

a search query performed on the system of information on the system or feedback of interaction by a user of the system with results returned from a search query performed on the system of information stored on the system, wherein the feedback of interaction by the user of the system is not transmitted to the remote search engine.

18. The system of claim 17, wherein training comprises a statistical analysis of the information obtained from the system, wherein statistical analysis comprises one of:

linear regression, Bayes classification, or Naive Bayes classification.

19. The system of claim 15, wherein information generated on the computing device comprises a context of the computing device at the time the search query was issued.

20. The system of claim 19, further comprising:

wherein the context includes an application from which the search query was issued.

21. The system of claim 16, further comprising:

anonymizing, by the system of the user, user interaction history with the remote search engine results, such that a particular user and/or particular computing device of the particular user is not identified with the anonymized user interaction history; and

transmitting the anonymized user interaction history with the remote search engine results to the remote search engine.