IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
UP-0878307
(2004-06-29)
|
등록번호 |
US-7720674
(2010-06-10)
|
발명자
/ 주소 |
- Kaiser, Matthias
- Klein, Jacob A
- Vogler, Hartmut
- Jiang, Shan
|
출원인 / 주소 |
|
대리인 / 주소 |
Finnegan, Henderson, Farabow, Garrett & Dunner LLP
|
인용정보 |
피인용 횟수 :
191 인용 특허 :
3 |
초록
▼
Methods and systems are provided for processing natural language queries. Such methods and systems may receive a natural language query from a user and generate corresponding semantic tokens. Information may be retrieved from a knowledge base using the semantic tokens. Methods and systems may levera
Methods and systems are provided for processing natural language queries. Such methods and systems may receive a natural language query from a user and generate corresponding semantic tokens. Information may be retrieved from a knowledge base using the semantic tokens. Methods and systems may leverage an interpretation module to process and analyze the retrieved information in order to determine an intention associated with the natural language query. Methods and systems may leverage an actuation module to provide results to the user, which may be based on the determined intention.
대표청구항
▼
What is claimed is: 1. A method for processing natural language queries comprising: obtaining a natural language query from a user; generating at least one semantic token from the natural language query; identifying data in a knowledge base using the at least one semantic token; interpreting the id
What is claimed is: 1. A method for processing natural language queries comprising: obtaining a natural language query from a user; generating at least one semantic token from the natural language query; identifying data in a knowledge base using the at least one semantic token; interpreting the identified data based on an intention associated with the user, wherein the intention is expressed in a personalized policy, and wherein interpreting the identified data comprises: recognizing an uncertainty in the natural language query, wherein the uncertainty comprises at least one of: a lack of identified data in the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the identified data including a plurality of conceptually similar elements that relate to a generated semantic token; and resolving the uncertainty based on the user intention, wherein resolving the uncertainty comprises: establishing the personalized policy for the user; applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty, the contextual data comprising data relating to the user's location and sensor data; receiving feedback from the user relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback; and actuating the interpreted data. 2. The method of claim 1, wherein obtaining the natural language query comprises obtaining at least one of an alphanumeric character, an audio signal, and a visual signal. 3. The method of claim 1, wherein identifying data in a knowledge base comprises searching a structured data archive using the semantic token. 4. The method of claim 1, wherein actuating the interpreted data comprises translating the interpreted data into at least one system-actionable command. 5. The method of claim 1, wherein actuating the interpreted data comprises providing the interpreted data to the user. 6. The method of claim 1, the method further comprising: identifying at least one term having a semantic relationship with the at least one semantic token. 7. The method of claim 6, wherein identifying data in a knowledge base comprises identifying data in a knowledge base using the at least one semantic token and the at least one semantically-related term. 8. The method of claim 7, wherein the at least one semantically-related term comprises a synonym. 9. The method of claim 1, wherein resolving the uncertainty further comprises establishing a dialog with the user. 10. The method of claim 1, wherein resolving the uncertainty further comprises applying a preset rule. 11. A method for processing natural language queries comprising: obtaining a natural language query from a user; generating at least one semantic token from the natural language query; identifying data in a knowledge base using the at least one semantic token; determining an intention associated with the user based on the identified data, wherein the intention is expressed in a personalized policy and, wherein determining the intention associated with the user comprises: recognizing an uncertainty in the natural language query, wherein the uncertainty comprises at least one of: a lack of identified data in the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the identified data including a plurality of conceptually similar elements that relate to a generated semantic token; and resolving the uncertainty, wherein resolving the uncertainty comprises: establishing the personalized policy for the user; applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty, the contextual data comprising data relating to the user's location and sensor data; receiving feedback from the user relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback; and providing information that is relevant to the natural language query to the user based on the determined intention. 12. The method of claim 11, wherein resolving the uncertainty further comprises establishing a dialog with the user. 13. The method of claim 11, wherein resolving the uncertainty further comprises applying a preset rule. 14. The method of claim 11, the method further comprising: identifying at least one term having a semantic relationship with the at least one semantic token generated from the natural language query. 15. The method of claim 14, wherein identifying data in a knowledge base comprises identifying data in a knowledge base using the at least one semantic token and the at least one semantically-related term. 16. The method of claim 15, wherein the at least one semantically-related term comprises a synonym. 17. A method for processing natural language queries comprising: obtaining a natural language query from a source; retrieving data that is potentially relevant to the natural language query from a knowledge base; determining an intention associated with the natural language query, wherein the intention is expressed in a personalized policy; and processing the potentially relevant data in accordance with the intention so as to identify actually relevant data from the potentially relevant data, wherein processing the potentially relevant data comprises: establishing the personalized policy for the user; applying the personalized policy in conjunction with contextual data obtained from the user to identify the actually relevant data from the potentially relevant data, the contextual data comprising data relating to the user's location and sensor data; receiving feedback from the user relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback; and providing the actually relevant data to the source. 18. The method of claim 17, wherein obtaining the natural language query comprises obtaining at least one of an alphanumeric character, an audio signal, and a visual signal from a user. 19. The method of claim 17, wherein retrieving the potentially relevant data comprises searching a structured data archive using at least one semantic token derived from the natural language query. 20. The method of claim 17, wherein determining the intention comprises interacting with a user to determine the intention. 21. The method of claim 17, wherein processing the potentially relevant data further comprises applying a preset rule in accordance with the intention. 22. A method for processing natural language queries, comprising: obtaining a natural language query from a user; generating at least one semantic token from the natural language query; identifying data in a knowledge base using the at least one semantic token; identifying an uncertainty in the natural language query, wherein the uncertainty comprises at least one of: a lack of identified data in the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the identified data including a plurality of conceptually similar elements that relate to a generated semantic token; determining an intention associated with the user based on the identified data, wherein the intention is expressed in a personalized policy; and resolving the identified uncertainty based on the determined intention, wherein resolving the uncertainty comprises: establishing the personalized policy for the user; applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty, the contextual data relating to the user's location and sensor data; receiving feedback from the user relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback. 23. The method of claim 22, wherein identifying an uncertainty comprises identifying that additional information is required from the user. 24. A system for processing natural language queries, comprising: means for obtaining a natural language query from a user; means for generating at least one semantic token from the natural language query; means for identifying data in a knowledge base using the at least one semantic token; means for determining an intention associated with the user based on the identified data, wherein the intention is expressed in a personalized policy, and wherein means for determining an intention associated with the user comprises: means for recognizing an uncertainty in the natural language query, wherein the uncertainty comprises at least one of: a lack of identified data in the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the identified data including a plurality of conceptually similar elements that relate to a generated semantic token; and means for resolving the uncertainty, the means for resolving the uncertainty comprising: means for establishing the personalized policy for the user; means for applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty the contextual data comprising data relating to the user's location and sensor data; means for receiving feedback from the user relating to the application of the personalized policy; and means for re-configuring the personalized policy based on the feedback; and means for providing information that is relevant to the natural language query to the user based on the determined intention. 25. A system for processing natural language queries, comprising: means for obtaining a natural language query from a user; means for generating at least one semantic token from the natural language query; means for identifying data in a knowledge base using the at least one semantic token; means for identifying an uncertainty in the natural language query, wherein the uncertainty comprises at least one of: a lack of identified data in the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the identified data including a plurality of conceptually similar elements that relate to a generated semantic token; means for determining an intention associated with the user based on the identified data, wherein the intention is expressed in a personalized policy; and means for resolving the identified uncertainty based on the determined intention, wherein the means for resolving the uncertainty comprises: means for establishing the personalized policy for the user; means for applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty, the contextual data comprising data relating to the user's location and sensor data; means for receiving feedback from the user relating to the application of the personalized policy; and means for re-configuring the personalized policy based on the feedback. 26. A natural language query processing system, comprising: an interface module configured to receive a natural language query; a tokenizing module configured to generate at least one semantic token based on the received natural language query; a searching module configured to retrieve information from a knowledge base using the at least one semantic token; an interpretation module configured to: identify an uncertainty associated with the natural language query, wherein the uncertainty comprises at least one of: a lack of retrieved information from the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the retrieved information including a plurality of conceptually similar elements that relate to a generated semantic token; and resolve the uncertainty, wherein resolving the uncertainty comprises: establishing a personalized policy, applying the personalized policy in conjunction with contextual data obtained from the user to resolve the uncertainty, the contextual data comprising data relating to the user's location and sensor data, receiving feedback relating to the application of the personalized policy, and re-configuring the personalized policy based on the feedback, and process the retrieved information so as to resolve the uncertainty based on an intention associated with the received natural language query, wherein the intention is expressed in the personalized policy; and an actuation module configured to translate the processed information into a system-actionable command. 27. The system of claim 26, wherein the interface module is configured to receive the natural language query from at least one of image capture device, an audio capture device, a keyboard, a mouse, a touch screen. 28. The system of claim 26, wherein the interface module is configured to receive the natural language query from a data processing system via a network. 29. The system of claim 26, wherein the tokenizing module is further configured to discard at least one element of the natural language query. 30. The system of claim 26, wherein the searching module is configured to retrieve the information from a structured data archive. 31. The system of claim 26, wherein the searching module is further configured to identify at least one term having a semantic relationship with the at least one semantic token. 32. The system of claim 31, wherein the searching module is configured to retrieve information from the knowledge base using the at least one semantic token and the semantically-related term. 33. The system of claim 32, wherein the semantically-related term comprises one of a synonym and a hypernym. 34. The system of claim 26, wherein the interpretation module is configured to resolve the uncertainty based on an intention associated with the received natural language query by retrieving information from a user. 35. The system of claim 26, wherein the interpretation module resolves the uncertainty by applying a preset rule. 36. A natural language query processing system, comprising: an interface module configured to receive a natural language query; a tokenizing module configured to generate at least one semantic token based on the received natural language query; a searching module configured to retrieve information from a knowledge base using the at least one semantic token; an interpretation module configured to: determine an intention associated with the received natural language query, wherein the intention is expressed in a personalized policy, and process the retrieved information in accordance with the intention, wherein the interpretation module processes the retrieved information by: establishing the personalized policy; applying the personalized policy in conjunction with contextual data obtained from the user to resolve an uncertainty in the natural language query, the contextual data comprising data relating to the user's location and sensor data; receiving feedback relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback; and an actuation module configured to provide the processed information to a user. 37. The system of claim 36, wherein the interface module is configured to receive the natural language query from at least one of image capture device, an audio capture device, a keyboard, a mouse, a touch screen. 38. The system of claim 36, wherein the interface module is configured to receive the natural language query from a data processing system via a network. 39. The system of claim 36, wherein the tokenizing module is further configured to discarding at least one element of the natural language query. 40. The system of claim 36, wherein the searching module is configured to retrieve the information from a structured data archive. 41. The system of claim 36, wherein the searching module is further configured to identify at least one term having a semantic relationship with the at least one semantic token. 42. The system of claim 41, wherein the searching module is configured to retrieve information from the knowledge base using the at least one semantic token and the semantically-related term. 43. The system of claim 42, wherein the semantically-related term comprises a synonym and a hypernym. 44. The system of claim 36, wherein the interpretation module is configured to determine the intention by communication with a user. 45. The system of claim 36, wherein the interpretation module determines the intention by: recognizing an uncertainty associated with the natural language query, wherein the uncertainty comprises at least one of: a lack of retrieved information from the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the retrieved information including a plurality of conceptually similar elements that relate to a generated semantic token; and retrieving clarifying data from a user in order to resolve the uncertainty. 46. The system of claim 45, wherein the interpretation module retrieves clarifying data from a user by establishing a dialog with the user via a data processing system. 47. The system of claim 36, wherein the interpretation module processes the retrieved information by applying a preset rule. 48. A computer-readable medium containing instructions for controlling a computer system coupled to a network to perform a method, the computer system having a processor for executing the instructions, the method comprising: obtaining a natural language query from a source; retrieving data that is potentially relevant to the natural language query from a knowledge base; determining an intention associated with the natural language query, wherein the intention is expressed in a personalized policy; and processing the potentially relevant data in accordance with the intention so as to separate the potentially relevant data into actually relevant data and actually irrelevant data, wherein the interpretation module processing the potentially relevant data comprises: establishing the personalized policy; applying the personalized policy in conjunction with contextual data obtained from the source to resolve an uncertainty in the natural language query, the contextual data comprising data relating to location and sensor data, wherein the uncertainty comprises at least one of: a lack of retrieved data from the knowledge base; the natural language query including a series of dependent commands without information associated with a sequence of the commands; and the retrieved data including a plurality of conceptually similar elements that relate to a generated semantic token; receiving feedback relating to the application of the personalized policy; and re-configuring the personalized policy based on the feedback; and providing the actually relevant data to the source.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.