DISAMBIGUATING ASSOCIATIONS

• US 20130198219A1
• Filed: 01/30/2012
• Published: 08/01/2013
• Est. Priority Date: 01/30/2012
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Cohen, Ira, Assulin, Ohad
• Granted Patent: US 8,949,264 B2
• Time in Patent Office: Days
• US Class Current: 707/760
• CPC Class Codes: G06F 16/243 Natural language query formulation

Abstract

Disclosed herein is a technique for disambiguating associations between one keyword and multiple attributes of a database model and for disambiguating associations between one attribute of a database model and multiple attribute types.

Claims

1. A system comprising:

instructions which, if executed, cause at least one processor to;

learn to translate natural language queries into database language queries using historical data associated with previously received natural language queries; and

respond to a new natural language query based on an analysis of the historical data, the historical data comprising;

associations between keywords contained in the previously received natural language queries and attributes of a database model, the associations being adaptable for disambiguating associations between one keyword and multiple attributes of the database model; and

a context free grammar, the context free grammar being adaptable for disambiguating associations between one attribute of the database model and multiple attribute types.

2. The system of claim 1, wherein each association between a keyword in the new natural language query and an attribute of the database model is further associated with a first probability, the first probability representing a likelihood that each association will translate into at least one expression of a database language query that produces a correct answer to the new natural language query.

3. The system of claim 2, wherein each sentence of the context free grammar comprises a formerly received natural language query in which each keyword therein that is associated with the attribute of the database model is replaced with a label representing a database attribute type, the database attribute type corresponding to the attribute of the database model associated with each keyword.

4. The system of claim 3, wherein each sentence of the context free grammar that contains at least one keyword also contained in the new natural language query is associated with a second probability, the second probability representing the likelihood that each sentence containing the at least one keyword will translate into at least one expression of the database language query that produces the correct answer to the new natural language query.

5. The system of claim 4, wherein the instructions, if executed, further cause the at least one processor to rank each association between the keyword in the new natural language query and the attribute of the database model based on the first probability associated with each association so as to assign a first ranking score to each association.

6. The system of claim 5, wherein the instructions, if executed, further cause the at least one processor to rank each sentence containing the at least one keyword also contained in the new natural language query based on the second probability associated with each sentence so as to assign a second ranking score to each sentence.

7. The system of claim 6, wherein the instructions, if executed, further cause the at least one processor to:

combine each first ranking score and each second ranking score so as to generate a final ranking score for each combination of first ranking score and second ranking score;

generate at least one database query such that expressions thereof include attributes of the database model and attribute types associated therewith that correspond to the final ranking score; and

execute the at least one database query in an order, the order coinciding with the final ranking score of each combination.

8. A non-transitory computer readable medium having instructions stored therein, which if executed, cause at least one processor to:

access associations between keywords contained in formerly received natural language queries and attributes of a database model;

access a context free grammar, each sentence of the context free grammar comprising a formerly received natural language query in which each keyword therein that is associated with an attribute of the database model is replaced with a label representing a database attribute type, the database attribute type corresponding to the attribute of the database model associated with each keyword;

disambiguate associations between one, keyword and multiple attributes of the database model;

disambiguate associations between one attribute of the database model and multiple database attribute types using the context free grammar; and

respond to a new natural language query.

9. The non-transitory computer readable medium of claim 8, wherein the instructions, if executed, further cause the at least one processor to assign a first probability to each association between a keyword in the new natural language query and the attribute of the database model, the first probability representing a likelihood that each association will translate into at least one expression of a database language query that produces a correct answer to the new natural language query.

10. The non-transitory computer readable medium of claim 9, wherein the instructions, if executed, further cause the at least one processor to assign a second probability to each sentence of the context free grammar that contains at least one keyword also contained in the new natural language query, the second probability representing the likelihood that each sentence containing the at least one keyword will translate into at least one expression of the database language query that produces the correct answer to the new natural language query.

11. The non-transitory computer readable medium of claim 10, wherein the instructions, if executed, further cause the at least one processor to rank each association between the keyword in the new natural language query and the attribute of the database model based on the first probability associated with each association so as to assign a first ranking score to each association.

12. The non-transitory computer, readable medium of claim 11, if executed, further cause the at least one processor to rank each sentence containing the at least one keyword also contained in the new natural language query based on the second probability associated with each sentence so as to assign a second ranking score to each sentence.

13. The non-transitory computer readable medium of claim 12, wherein the instructions, if executed, further cause the at least one processor to:

combine each first ranking score and each second ranking score so as to generate a final ranking score for each combination of first ranking score and second ranking score;

generate at least one database query such that expressions thereof include attributes of the database model and attribute types associated therewith that correspond to the final ranking score; and

execute the at least one database query in an order, the order coinciding with the final ranking score of each combination.

14. A method comprising:

disambiguating associations between one keyword and multiple attributes of a database model;

disambiguating associations between one attribute of the database model and multiple database attribute types using a context free grammar; and

responding to a new natural language query.

15. The method of claim 14, further comprising:

accessing associations between keywords contained in formerly received natural language queries and attributes of the database model; and

accessing the context free grammar, each sentence of the context free grammar comprising a formerly received natural language query in which each keyword therein that is associated with an attribute of the database model is replaced with a label representing the database attribute type, the database attribute type corresponding to the attribute of the database model associated with each keyword.

16. The method of claim 14, further comprising:

assigning a first probability to each association between a keyword in the new natural language query and an attribute of the database model, the first probability representing a likelihood that each association will translate into at least one expression of a database language query that produces a correct answer to the new natural language query.

17. The method of claim 16, further comprising:

assigning a second probability to each sentence of the context free grammar that contains at least one keyword also contained in the new natural language query, the second probability representing the likelihood that each sentence containing the at least one keyword will translate into at least one expression of the database language query that produces the correct answer to the new natural language query.

18. The method of claim 17, further comprising:

ranking each association between the keyword in the new natural language query and the attribute of the database model based on the first probability associated with each association so as to assign a first ranking score to each association.

19. The method of claim 18, further comprising:

ranking each sentence containing the at least one keyword also contained in the new natural language query based on the second probability associated with each sentence so as to assign a second ranking score to each sentence.

20. The method of claim 19, further comprising

combine each first ranking score and each second ranking score so as to generate a final ranking score for each combination of first ranking score and second ranking score;

generating at least one database query such that expressions thereof include attributes of the database model and attribute types associated therewith that correspond to the final ranking score; and

executing the at least one database query in an order, the order coinciding with the final ranking score of each combination.