KNOWLEDGE REPOSITORY

• US 20070055656A1
• Filed: 07/21/2006
• Published: 03/08/2007
• Est. Priority Date: 08/01/2005
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Evi Technologies Ltd. (Amazon.com, Inc.)
• Inventors: TUNSTALL-PEDOE, WILLIAM
• Granted Patent: US 8,666,928 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/532 Query formulation, e.g. graphical querying
• CPC Class Codes: G06F 40/279 Recognition of textual entities
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06N 5/00 Computing arrangements using knowledge-based models

Abstract

A knowledge storage system is described. A specific embodiment is a computer system comprising a knowledge base of general knowledge in structured form which can be added to and queried by untrained users. Various embodiments include the facility for remote computers to access the knowledge stored in the system, natural language questions to be answered, profile screens giving general knowledge about an object in the system, and methods for distinguishing between reliable and unreliable facts.

Claims

1. A knowledge representation system, comprising:

a data store having a knowledge base stored therein comprising first knowledge represented in a structured, machine-readable format which encodes meaning; and

at least one computing device operable to;

add second knowledge to the knowledge base, the second knowledge being generated with reference to input from a plurality of users which is not in the structured, machine-readable format, at least some of the input from the users being in a natural language;

generate third knowledge not represented in the knowledge base by inferring the third knowledge from at least one of the first knowledge and the second knowledge; and

respond to queries using at least one of the first knowledge, the second knowledge, and the third knowledge.

2. The system of claim 1 wherein the at least one computing device is further operable to translate natural language questions received from selected ones of the users to the queries.

3. The system of claim 1 wherein the at least one computing device is operable to add the third knowledge to the knowledge base by capturing responses from a first one of the users to a sequence of natural language questions, aggregating the responses, and translating the aggregated responses to the structured, machine-readable format.

4. The system of claim 3 wherein the at least one computing device is further operable to present a natural language translation of at least some of the aggregated responses to the first user for confirmation by the first user.

5. The system of claim 1 wherein the second knowledge comprises a plurality of assertions.

6. The system of claim 5 wherein selected ones of the assertions have an associated source identifier identifying a corresponding one of the users as a source of the associated assertion.

7. The system of claim 5 wherein the at least one computing device is further operable to publish a first one of the assertions to selected ones of the users prior to general publication of the first assertion in the knowledge base, and to add the first assertion to the knowledge base in response to endorsement of the first assertion by the selected users.

8. The system of claim 5 wherein the at least one computing device is further operable to inhibit use of a first one of the assertions in the knowledge base in response to contradiction of the first assertion by at least one of the users.

9. The system of claim 5 wherein the at least one computing device is further operable to analyze a first one of the assertions with reference to the knowledge base, and to inhibit use of the first assertion in the knowledge base where the first assertion is in semantic conflict with the knowledge base.

10. The system of claim 9 wherein the at least one computing device is further operable to permit use of the first assertion in the knowledge base in response to changes in the knowledge representation system which remove the semantic conflict.

11. The system of claim 5 wherein the at least one computing device is further operable to analyze a first one of the assertions, and to inhibit use of the first assertion in the knowledge base where the first assertion corresponds to a prohibited assertion type.

12. The system of claim 1 wherein the at least one computing device is operable to respond to a knowledge request corresponding to a first one of the queries using at least one of the first knowledge, the second knowledge, the third knowledge, and search results comprising a plurality of natural language documents retrieved using a conventional search engine.

13. The system of claim 1 wherein the at least one computing device is further operable to receive a first one of the queries, the first query being generated by an interface operating in conjunction with an application on a remote computing device.

14. The system of claim 1 wherein the at least one computing device is further operable to authenticate an entity attempting to interact with the knowledge base by attempting to uniquely identify the entity with reference to the knowledge base, and to facilitate addition of information representing the entity to the knowledge base where the entity is not already represented in the knowledge base.

15. The system of claim 14 wherein the at least one computing device is further operable to verify that the entity corresponds to the information representing the entity in the knowledge base by determining a correlation between the entity and the information using evidence derived using at least one authentication technique.

16. The system of claim 1 wherein the at least one computing device is further operable to provide incentives to the users for providing the input with reference to which the second knowledge is generated.

17. The system of claim 1 wherein the first knowledge includes information relating to real world objects, and wherein the at least one computing device is further operable to present profiles identifying the real world objects, each profile corresponding to at least one of the real world objects and including human-perceivable facts relating to the at least one real world object.

18. A computer-implemented method for facilitating addition to a knowledge base, the knowledge base comprising first knowledge represented in a structured, machine-readable format which encodes meaning, the method comprising:

providing at least one interface by which a first user may enter information which is not in the structured, machine-readable format, at least some of the information being in a natural language, the at least one interface being operable to transmit the information to at least one remote computing device for generation of second knowledge represented in the machine-readable format for addition to the knowledge base; and

presenting responses to knowledge requests using at least one of the first knowledge, the second knowledge, and third knowledge not represented in the knowledge base and inferred from at least one of the first knowledge and the second knowledge.

19. The method of claim 18 wherein the at least one interface is operable to capture responses from the first user to a sequence of natural language questions.

20. The method of claim 19 further comprising presenting a natural language translation of an aggregation of at least some of the responses to the first user for confirmation by the first user.

21. The method of claim 18 further comprising facilitating input by at least one other user regarding credibility of the information.

22. The method of claim 18 wherein the responses to selected ones of the knowledge requests include search results comprising a plurality of natural language documents retrieved using a conventional search engine

23. The method of claim 18 further comprising providing at least one incentive to the first user for entering the information.

24. The method of claim 18 wherein the first knowledge includes information relating to real world objects, the method further comprising presenting profiles identifying the real world objects, each profile corresponding to at least one of the real world objects and including human-perceivable facts relating to the at least one real world object.

25. A computer program product comprising at least one computer-readable medium having computer program instructions stored therein which are operable to cause a least one computer to perform the method of claim 18.

26. A computing system in a network, the system comprising a knowledge repository comprising first knowledge represented in a structured, machine-readable format operable to store information about any entity that can be denoted in a natural language, the system further comprising at least one computing device operable to facilitate addition of second knowledge to the knowledge repository by collecting input from a plurality of users via the network using natural language requests, and translating the input to the machine-readable format.

27. The system of claim 26 wherein the second knowledge comprises a plurality of assertions, and wherein the at least one computing device is further operable to inhibit use of selected ones of the assertions in the knowledge repository.

28. The system of claim 27 wherein the at least one computing device is operable to inhibit use of the selected assertions in response to contradiction of the selected assertions by selected ones of the users.

29. The system of claim 27 wherein the at least one computing device is operable to inhibit use of the selected assertions where the selected assertions are in semantic conflict with the first knowledge.

30. The system of claim 27 wherein the at least one computing device is operable to inhibit use of the selected assertions where the selected assertions correspond to a prohibited assertion type.

31. A computing system in a network, the system comprising a knowledge repository comprising first knowledge represented in a structured, machine-readable format operable to store information about any entity that can be denoted in a natural language, the system further comprising at least one computing device operable to facilitate addition of second knowledge to the knowledge repository by a plurality of users without requiring knowledge of the machine-readable format by the users.

32. The system of claim 31 wherein the second knowledge comprises a plurality of assertions, and wherein the at least one computing device is further operable to inhibit use of selected ones of the assertions in the knowledge repository.

33. The system of claim 32 wherein the at least one computing device is operable to inhibit use of the selected assertions in response to contradiction of the selected assertions by selected ones of the users.

34. The system of claim 32 wherein the at least one computing device is operable to inhibit use of the selected assertions where the selected assertions are in semantic conflict with the first knowledge.

35. The system of claim 32 wherein the at least one computing device is operable to inhibit use of the selected assertions where the selected assertions correspond to a prohibited assertion type.

36. A computer-implemented method for responding to knowledge requests, comprising:

presenting a first natural language response to a first knowledge request, the first natural language response being derived from knowledge represented in a structured, machine-readable format operable to store information about any entity that can be denoted in a natural language; and

presenting search results in response to a second knowledge request where a second natural language response derived from the knowledge is not available, the search results comprising a plurality of natural language documents identified using a conventional search engine.

37. A computer-implemented method for responding to a knowledge request, comprising:

presenting a natural language response to the knowledge request, the natural language response being derived from knowledge represented in a structured, machine-readable format operable to store information about any entity that can be denoted in a natural language; and

presenting search results in conjunction with the natural language response, the search results comprising a plurality of natural language documents retrieved using a conventional search engine.

38. A computer-implemented method for facilitating addition of a first entity to a knowledge base by a first human user, the method comprising:

facilitating identification by the first human user of a class to which the first entity belongs;

facilitating generation by the first human user of at least one first natural language string denoting the first entity;

facilitating generation by the first human user of at least one second natural language string corresponding to the first entity, the at least one second natural language string being specified to facilitate unique recognition of the first entity by humans; and

facilitating transmission of data representing the class and the first and second natural language strings for storage in a knowledge base in association with an identifier uniquely identifying the first entity within the knowledge base.

39. The method of claim 38 further comprising:

presenting at least one third natural language string to the first human user, the at least one third natural language string corresponding to a second entity represented in the knowledge base and being specified to facilitate unique recognition of the second entity by humans; and

facilitating verification by the first human user that the first entity is distinct from the second entity with reference to the at least one third natural language string.