Knowledge repository

• US 8,666,928 B2
• Filed: 07/21/2006
• Issued: 03/04/2014
• 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
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): BUSS, BENJAMIN J
• Application Number:
  US11/459,202
• Publication Number:
  US 20070055656A1
• Time in Patent Office: 2,783 Days
• Field of Search: 706/55, 706/45, 707/100, 707/101, 707/102, 707/103.R, 707/603, 707/690, 707/792, 707/794, 707/802, 707/805, 707/811, 707/955, 707/999.102, 707/999.103
• US Class Current: 706/55
• 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 is distinct from natural language, the structured, machine-readable format primarily comprising assertions of named relationships between pairs of named entities; and

at least one computing device operable to;

add second knowledge in the structured, machine-readable format to the knowledge base, the second knowledge being generated with reference to input from a plurality of untrained, general internet users which is not in the structured, machine-readable format, at least some of the input from the users being in a particular natural language, wherein generating the second knowledge includes translating the input from the users to factual assertions compatible with the structured, machine-readable format using a plurality of translation templates, each of the translation templates including a respective predetermined pattern for matching against one or more natural language strings included in the input from the users, and wherein adding the second knowledge to the knowledge base includes determining whether the second knowledge is semantically contradicted by the first knowledge to promote consistency of the first and second knowledge across the knowledge base;

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

respond to queries using one or more of the first knowledge, the second knowledge, or 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 one or more of the first knowledge, the second knowledge, the third knowledge, or 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 and is distinct from natural language, the structured, machine-readable format primarily comprising assertions of named relationships between pairs of named entities, the method comprising:

providing at least one interface by which a first untrained, general internet user may enter information which is not in the structured, machine-readable format, at least some of the information being in a particular 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, wherein generation of the second knowledge includes translating the input from the users to factual assertions compatible with the structured, machine-readable format using a plurality of translation templates, each of the translation templates including a respective predetermined pattern for matching against one or more natural language strings included in the input from the users, and wherein addition of the second knowledge to the knowledge base includes determining whether the second knowledge is semantically contradicted by the first knowledge to promote consistency of the first and second knowledge across the knowledge base; and

presenting responses to knowledge requests using one or more of the first knowledge, the second knowledge, or third knowledge not represented in the knowledge base, the third knowledge being inferred from one or more of the first knowledge or 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 that is distinct from natural language and is operable to store information about any entity that can be denoted in natural language, the structured, machine-readable format primarily comprising assertions of named relationships between pairs of named entities, 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 untrained, general internet users via the network using natural language requests, and translating the input to the machine-readable format using a plurality of translation templates, each of the translation templates including a respective predetermined pattern for matching against one or more natural language strings included in the input, wherein addition of the second knowledge to the knowledge base includes determining whether the second knowledge is semantically contradicted by the first knowledge to promote consistency of the first and second knowledge across the knowledge base.

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 that is distinct from natural language and is operable to store information about any entity that can be denoted in natural language, the structured, machine-readable format primarily comprising assertions of named relationships between pairs of named entities, the system further comprising at least one computing device operable to facilitate addition of second knowledge in the machine-readable format to the knowledge repository by a plurality of untrained, general internet users without requiring knowledge of the machine-readable format by the users, wherein generation of the second knowledge includes translating input from the users to factual assertions compatible with the structured, machine-readable format using a plurality of translation templates, each of the translation templates including a respective predetermined pattern for matching against one or more natural language strings included in the input from the users, and wherein addition of the second knowledge to the knowledge base includes determining whether the second knowledge is semantically contradicted by the first knowledge to promote consistency of the first and second knowledge across the knowledge base.

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.