RESPONDING TO NATURAL LANGUAGE QUERIES

• US 20140095145A1
• Filed: 09/28/2012
• Published: 04/03/2014
• Est. Priority Date: 09/28/2012
• Status: Active Grant

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Assulin, Ohad, Cohen, Ira, Mordechai, Eli, Shor, Boaz, Sade, Alon
• Granted Patent: US 9,411,803 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30

Abstract

Disclosed herein are a system, non-transitory computer-readable medium, and method for responding to natural language queries. Keywords likely to appear in a natural language query are determined and each likely keyword is associated with a module. A response to a natural language query comprises information generated by each module associated with a likely keyword appearing in the natural language query.

Claims

1. A system comprising:

at least one processor;

an indexer which, if executed, causes the at least one processor to determine which keywords are likely to appear in a natural language query and to associate each likely keyword with a module of a plurality of modules;

a query translator which, if executed, causes the at least one processor to determine whether at least one of the likely keywords determined by the indexer appears in a received natural language query; and

a results generator which, if executed, causes the at least one processor to respond to the received natural language query with information generated by each module associated with a likely keyword appearing in the received natural language query.

2. The system of claim 1, further comprising a language learner which, if executed, causes the at least one processor to alter an association between a likely keyword and a module of the plurality of modules, when it is determined that the altered association is more likely to trigger an accurate response to a natural language query.

3. The system of claim 1, wherein the results generator, if executed, further causes the at least one processor to rank the information returned by each module based on a probability that the information is a correct response to the received natural language query.

4. The system of claim 3, wherein the probability is at least partially based on a number of associations between each module and likely keywords appearing in the received natural language query.

5. The system of claim 3, wherein the probability is at least partially based on an amount of information returned by each module.

6. The system of claim 1, wherein the indexer, if executed, further causes the at least one processor to determine a keyword likely to appear in a natural language query from source code text of a module of the plurality of modules.

7. The system of claim 1, wherein the indexer, if executed, further causes the at least one processor to associate each likely keyword with an attribute of a module of the plurality of modules based on metadata associated with each attribute.

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

associate each potential keyword that could appear in a natural language query of a plurality of possible natural language queries with a module in a plurality of modules likely to provide accurate answers to the plurality of possible natural language queries;

alter an association between a potential keyword and a module in the plurality of modules, when it is determined that the altered association is more likely to trigger an accurate response to a natural language query of the plurality of possible natural language queries;

determine whether at least one potential keyword appears in a received natural language query; and

respond to the received natural language query with data produced by each module associated with a potential keyword that appears in the received natural language query.

9. The non-transitory computer-readable medium of claim 8, wherein the instructions stored in the computer-readable medium, when executed, further causes the at least one processor to rank the data returned by each module based on a probability that the data is a correct response to the received natural language query.

10. The non-transitory computer-readable medium of claim 9, wherein the probability is at least partially based on a number of associations between each module and potential keywords appearing in the received natural language query.

11. The non-transitory computer-readable medium of claim 9, wherein the probability is at least partially based on an amount of information returned by each module.

12. The non-transitory computer-readable medium of claim 8, wherein the instructions stored in the computer-readable medium, when executed, further causes the at least one processor to derive a potential keyword from source code text of a module in the plurality of modules.

13. The non-transitory computer-readable medium of claim 8, wherein the instructions stored in the computer-readable medium, when executed, further causes the at least one processor to associate each potential key word with an attribute of a module in the plurality of modules based on metadata associated with each attribute.

14. A method comprising:

determining, using at least one processor, keywords that are likely to appear in a plurality of possible natural language queries such that some keywords are determined using source code text of modules likely to provide accurate answers to the plurality of possible natural language queries;

associating, using the at least one processor, each determined keyword with a module of the modules likely to provide an accurate answer to the plurality of natural language queries;

changing, using the at least one processor, an association between a determined keyword and a module, if it is determined that changing the association is more likely to trigger an accurate response to a natural language query of the plurality of natural language queries;

determining, using the at least one processor, whether at least one of the determined keywords appears in a received natural language query; and

generating, using the at least one processor, a response to the received natural language query with data produced by each module associated with a determined keyword appearing in the received natural language query.

15. The method of claim 14, further comprising ranking, using the at least one processor, the data returned by each module based on a probability that the data is a correct response to the received natural language query.

16. The method of claim 15, wherein the probability is at least partially based on a number of associations between each module and determined keywords appearing in the received natural language query.

17. The method of claim 15, wherein the probability is at least partially based on an amount of information returned by each module.

18. The method of claim 14, wherein associating each determined keyword with a module comprises associating, using the at least one processor, each determined keyword with an attribute of a module included in the modules likely to provide accurate answers to the plurality of possible natural language queries based on metadata associated with each attribute.