Disambiguation of Entities

• US 20120096033A1
• Filed: 10/14/2010
• Published: 04/19/2012
• Est. Priority Date: 10/14/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: GOVANI, TABREEZ, SHAH, SARTHAK DEEPAK, HOSSEINI, FARID
• Granted Patent: US 9,418,155 B2
• Time in Patent Office: Days
• US Class Current: 707/780
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9532 Query formulation

Abstract

Methods, systems, algorithms, and media are provided for disambiguating entities present in a received search query. Lists of categories from semi-structured data from external sites as well as internal sources are used to detect if ambiguity exists in an entity within the search query. Multiple senses or categories of the ambiguous entity are determined by ascertaining the primary intent of an entity extracted from a main term of a document. The probability of each sense is calculated by computing a total amount of traffic received for each of the senses of the ambiguous entity. The sense with the highest amount of computed traffic is the most probable determined sense.

Claims

1. A computer-implemented method of disambiguating entities using a computing system having processor, memory, and data storage subsystems, the computer-implemented method comprising:

receiving a user input search query;

detecting if ambiguity exists in an entity within the search query;

determining multiple senses that exist within the detected ambiguous entity;

calculating a probability of each determined sense of the detected ambiguous entity; and

returning search results for a most probable determined sense of the detected ambiguous entity.

2. The computer-implemented method of claim 1, wherein said detecting utilizes lists of categories from semi-structured data.

3. The computer-implemented method of claim 1, wherein each of the determined multiple senses are determined by extracting a main term from a document to ascertain its primary intent.

4. The computer-implemented method of claim 1, wherein the calculating comprises computing a total amount of traffic received for each determined sense of the detected ambiguous entity.

5. The computer-implemented method of claim 4, wherein the most probable determined sense comprises the determined sense with a highest amount of computed traffic.

6. The computer-implemented method of claim 1, wherein the returned search results comprise content that is proportional to the calculated probability of each of the determined multiple senses.

7. The computer-implemented method of claim 1, wherein the returned search results comprise content for all determined multiple senses with a probability above a minimum probability threshold level and omits content for all determined multiple senses with a probability below the minimum probability threshold level.

8. The computer-implemented method of claim 1, wherein the returned search results comprise content for the most probable determined sense and a module comprising links to all other determined multiple senses of the detected ambiguous entity.

9. The computer-implemented method of claim 8, wherein the module further comprises definitions of the other determined multiple senses.

10. One or more computer-readable storage media containing computer readable instructions for an algorithm embodied thereon that, when executed by a computing device, perform steps for disambiguating entities, the algorithm comprising:

detecting if ambiguity exists for an entity obtained from a search query;

determining senses that exist within the detected ambiguous entity;

computing a total amount of network traffic for each of the determined senses; and

calculating a probability for each of the determined senses of the detected ambiguous entity via the processor of the computing device.

11. The one or more computer-readable storage media of claim 10, wherein the calculating comprises dividing the computed network traffic for each of the determined senses by a total amount of computed network traffic for all determined senses of the detected ambiguous entity.

12. The one or more computer-readable storage media of claim 10, wherein the calculated probability of each of the determined senses utilizes web browser data.

13. The one or more computer-readable storage media of claim 12, wherein the probability of each of the determined senses is calculated using web browser page views and dwell time for each of the web browser page views.

14. The one or more computer-readable storage media of claim 10, wherein the ambiguity is detected when an entity appears in multiple categorized lists.

15. A computer-implemented method of detecting an ambiguous search query using a computing system having processor, memory, and data storage subsystems, the computer-implemented method comprising:

receiving a search query from a user input via an interconnected computing network of the computing system;

identifying an ambiguous term in the search query by utilizing lists of categories from semi-structured data containing the ambiguous term;

inferring categories of the identified ambiguous term via extraction on the semi-structured data;

determining a probability for each inferred category of the identified ambiguous term from web browser data via the processor of the computing system; and

returning search results representing a most probable determined category of the identified ambiguous term to a user via a graphical user interface of the computing system.

16. The computer-implemented method of claim 15, wherein the web browser data comprises a number of page views and dwell time for the page views for each of the determined categories.

17. The computer-implemented method of claim 16, wherein the determining comprises measuring an amount of computing traffic for the page views for each of the determined categories.

18. The computer-implemented method of claim 15, wherein the returned search results comprise content for the most probable category which exceeds a maximum probability threshold.

19. The computer-implemented method of claim 15, wherein the returned search results comprise content for all of the inferred categories with a probability above a minimum probability threshold level and omits content for all of the inferred categories with a probability below the minimum probability threshold level.

20. The computer-implemented method of claim 15, wherein the returned search results comprise content that is proportional to the determined probability of each of the inferred categories.