Knowledge web

• US 20050086188A1
• Filed: 04/10/2002
• Published: 04/21/2005
• Est. Priority Date: 04/11/2001
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Metaweb Technologies, Inc. (Alphabet Inc.)
• Inventors: Hillis, Daniel W., Ferren, Bran
• Granted Patent: US 7,502,770 B2
• Time in Patent Office: Days
• US Class Current: 706/50
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G09B 19/00 Teaching not covered by other main groups of this subclass teaching or practice apparatus for gun-aiming or gun-laying F41G3/26
• CPC Class Codes: G09B 5/00 Electrically-operated educational appliances working with questions and answers G09B7/00; simulators G09B9/00; advertising or displaying in general G09F

Abstract

A system and method for organizing knowledge in such a way that humans can find knowledge, learn from it, and add to it as needed is disclosed. The exemplary system has four components: a knowledge base (215), a learning model (21) and an associated tutor (13), a set or user tools (31, 32), and a backend system (33). The invention also preferably comprises a set of application programming interfaces (APIs) that allow these components to work together, so that other people can create their own versions of each of the components. In the knowledge web a community of people with knowledge to share put knowledge in the database using the user tools. The knowledge may be in the form of documents or other media, or it may be a descriptor of a book or other physical source. Each piece of knowledge is associated with various types of meta-knowledge about what the knowledge is for, what form it is in, and so on.

Claims

1. A system for organizing knowledge in such a way that users can find knowledge, learn from it, and add to it as needed, comprising:

a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said knowledge, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary by users, and connections between said knowledge and other units of knowledge;

a set of user tools comprising one or more tools for entering said knowledge, said meta-knowledge, and said one or more annotations into said knowledge base; and

a user learning model comprising any of information on a user'"'"'s needs, capabilities, knowledge, and preferences, said meta-knowledge stored in said knowledge base, and generalized knowledge about how people learn.

2. The system of claim 1, further comprising:

a tutor that takes user preferences into account, and that optionally uses said meta-knowledge and said user learning model to plan what information is presented to said user and how said information is presented.

3. A method for organizing knowledge in such a way that users can find it, learn from it, and add to it as needed, comprising any of the steps of:

finding chains of explanations that connect concepts a user needs to learn to what said user already knows;

showing said user a map of what said user needs to learn;

choosing explanatory paths that match said user'"'"'s preferred style of learning, optionally including any of enough side paths, interesting examples, multimedia documents, and related curiosities to match said user'"'"'s level of interest;

whenever possible, following explanatory paths laid down by great teachers;

if an explanation does not work, and consistently raises a particular type of question, then recording this information in a database where it can be used in planning paths of other students;

once said user has learned said concepts, updating said database to indicate that said user has recently learned said concepts;

reinforcing said user'"'"'s learning by finding connections that tie said concepts together;

keeping track of changes in said user'"'"'s preferred style of learning and pre-existing knowledge;

when an explanation does not work, trying another approach;

said user performing any of probing further, requesting examples, and living said database explicit feedback; and

said database using feedback to adjust a lesson, and thereby learning more about said user.

4. A system for organizing knowledge in such a way that users can find knowledge, learn from it, and add to it as needed, comprising:

a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said knowledge, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary by users, and connections between said knowledge and other units of knowledge;

a viewing tool for accessing said knowledge;

a tutor for maintaining a model of a user, and for finding useful knowledge to present to said user;

an authoring tool for adding knowledge into said knowledge base; and

a backend system comprising a database that stores and manages said knowledge base.

5. The system of claim 4, further comprising:

a knowledge representation scheme for organizing said knowledge base into explanations, topics, explanatory paths, and meta-knowledge representing information about and relationships between said knowledge in said knowledge base.

6. The system of claim 5, wherein an explanation comprises a piece of content that is linked with topics;

wherein every explanation has links to topics that it explains;

wherein topics and explanations are linked by said paths; and

wherein said meta-knowledge comprises one or more annotations to said explanations, topics, and.

7. The system of claim 5, wherein explanations also have links to topics that represent prerequisite knowledge.

8. The system of claim 5, wherein a topic is a cluster of concepts that a user might want to learn together;

and wherein topics optionally have multiple subtopics included within them.

9. The system of claim 8, said topics comprising a testable unit of knowledge that contains no subtopics, that represents a single idea or a fact, so that a single question may be created to test whether said user knows said testable unit of knowledge or not.

10. The system of claim 5, wherein a path is a way of describing a sequence of explanations and queries, with possible branch points;

wherein paths encode information about ways to learn a topic;

wherein a path is connected to topics that it explains and topics that it depends on as prerequisites.

11. The system of claim 10, wherein a path can contain branch points that are based on answers to queries.

12. The system of claim 10, wherein a path may contain additional information indicating how a sequence is presented.

13. The system of claim 5, wherein annotations are associated with explanations, topics, paths, or other annotations.

14. The system of claim 13, wherein annotations do not modify an author'"'"'s content, but add to it.

15. A system for organizing knowledge in such a way that users can find knowledge, learn from it, and add to it as needed, comprising:

a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said knowledge, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary by users, and connections between said knowledge and other units of knowledge;

a viewing tool for accessing said knowledge, wherein said viewing tool supports choosing topics that a user wants to learn about, viewing explanations provided to said user as a sequence of presentations, and annotating;

a tutor for maintaining a user learning model, and for finding useful knowledge to present to said user;

an authoring tool for enabling an author to add knowledge into said knowledge base; and

a backend system comprising a database system that stores and manages said knowledge base.

16. The system of claim 15, wherein said choosing is accomplished by naming a topic via entering a word or phrase into a topic-search engine;

wherein said viewing tool then displays a map of an area of a topic space said user selects, showing what said user already knows and what is knowable.

17. The system of claim 15, further comprising:

a topic map in which a space of topics and subtopics is illustrated as an n-dimensional landscape, with optional border, landmarks and links showing relationships between topics;

wherein an optional coloring scheme shows said user'"'"'s prior knowledge and relative importance of said topic.

18. The system of claim 15, wherein said topic map shows paths that said user has traveled before and paths that others have traveled before;

and wherein said viewing tool optionally allows said user to move through topic space by panning, zooming, or leaping from topic to related topic;

wherein said user can optionally zoom into relevant topics, look at their subtopics and mark things that are of interest, or that are already known.

19. The system of claim 15, wherein collecting a simple poll via an optional voting scheme indicates whether a user liked an explanation;

wherein an encryption scheme optionally hides a user'"'"'s identity, and optionally guarantees that a user can vote only once.

20. The system of claim 19, wherein said simple poll comprises a voting scheme comprising a user operated forward and back selection mechanism.

21. The system of claim 15, wherein users can make annotations explicitly which contain descriptive information about how an annotation is related;

wherein an annotation of this type is treated as content, and must have an author who declares authorship of said annotation, and wherein only said author can modify or delete said annotation.

22. The system of claim 15, wherein said tutor keeps track of data about said user, and data about said knowledge base;

and wherein all user-specific data are optionally private and inaccessible to others.

23. The system of claim 22, wherein said tutor planner uses awareness of any of said user'"'"'s age, language preferences, and reading level to choose and sort explanations.

24. The system of claim 22, wherein said tutor follows explicit paths that have been laid down by teachers, taking advantage of their suggestions about how to present said knowledge.

25. The system of claim 22, further comprising:

at least one explanation having properties which comprise any of annotation for adding knowledge into said knowledge base, and registration by which knowledge is declared to exist as part of said knowledge base.

26. The system of claim 25, wherein registration is accomplished by submitting knowledge to a registration server;

wherein before content can be registered specific meta-knowledge must be added and various options specified, which optionally include any of an author, creation date, URL identifying where said knowledge is stored, a list of topics an explanation explains and type of information specifying language, media type, explanation category, and a set of prerequisite topics, with a required level of expertise specified for each topic.

27. The system of claim 15, further comprising:

a mechanism for helping to find topics corresponding to an explanation;

wherein an author specifies a topic to which an explanation applies using a topic chooser;

wherein an authoring tool then presents said author with a list of specific topics and/or subtopics, sorted according to how well they match said explanation;

wherein said author may choose one or more of said topics and/or subtopics, and optionally narrow down a range to specific testable units of knowledge.

28. The system of claim 15, further comprising:

one or more registries handled by a registration server, which registries may comprise any of a pen name registry, a content registry and a topic registry.

29. The system of claim 28, wherein an author must register content under a pen name, wherein said pen name must itself be registered with said pen name registry;

wherein a pen name may be a real name or an alias;

wherein pen names are unique identifiers;

wherein said registration server does not register a same pen name to two different people;

wherein a pen name may be registered anonymously, in which case it is identified as such;

wherein a single author may have more than one pen name; and

wherein each pen name optionally has an associated password, which is used to verify said author'"'"'s identity.

30. The system of claim 28, wherein said registration server keeps a registry of all content in said knowledge base, including any explanations, queries, paths, and annotations.

31. The system of claim 28, wherein said registration server keeps track of where information is, said author'"'"'s pen name, and when said information was registered.

32. The system of claim 28, wherein said registration server keeps a hash-coded signature of registered content to verify that said content has not been changed.

33. The system of claim 28, wherein when an author registers content, said author must affirm that said author either owns said content, or has a right to publish said content;

wherein if there are access restrictions on said content, said registration can specify a permission server that is empowered to negotiate access.

34. The system of claim 28, wherein said topic registry keeps track of all topics and testable units of knowledge.

35. An apparatus for storing, organizing, and sharing a very large amount of loosely structured data among a large and diverse group of users, comprising:

a database that can be accessed and modified by thousands of users concurrently;

a distributed registry for keeping track of where and how said data and associated metadata are stored;

wherein data objects are represented as nodes of a labeled graph, and said associated metadata are represented by labeled links connecting said nodes;

wherein said nodes represent data of different types and in different formats, including text, image, sound, video, and structured data.

36. The apparatus of claim 35, further comprising:

a metaweb server which has access to a user'"'"'s security profile, and access to said distributed registry;

wherein said metaweb server obtains a location of data requested by said user, retrieves said data from data server which themselves maintain rules regarding access to retrieve and update data from said data servers, and assembles said data into coherent data structures.

37. The apparatus of claim 35, wherein each object is registered in said registry;

wherein said registry keeps track of where data and metadata associated with an object are stored;

wherein every data object has a unique index which is used to access the registry;

wherein said apparatus retrieves said data object from said index by assembling components of data and metadata from various storage devices across a network.

38. The apparatus of claim 35, further comprising:

a hash module, wherein when a data object is registered, said data object type and content are used to generate a fast, unique hash value, which is used as an index into said registry;

wherein said index is used to identify said data object in said registry and is used as an index in a registry hash table comprising a plurality of said hash values.

39. The apparatus of claim 35, wherein said registry further comprises:

a hash to an index to keep track of a descriptor for allowing a data object to be constructed from data stored in multiple locations.

40. The apparatus of claim 35, wherein said registry further comprises:

a hash to an index to keep track of a descriptor for allowing a data object to be stored redundantly in multiple locations.

41. The apparatus of claim 35, wherein said registry further comprises:

a hash to an index to keep track of a descriptor for allowing a data object to be stored in a different location than that of associated metadata.

42. The apparatus of claim 35, wherein said registry further comprises:

a hash to an index to keep track of a descriptor of a node of a labeled graph stored in multiple locations.

43. The apparatus of claim 35, wherein said registry further comprises:

an access process comprising a fast hash and a cryptographically strong signature.

44. The apparatus of claim 35, wherein said registry further comprises:

an hierarchical distributed cache.

45. The apparatus of claim 35, further comprising:

a metaweb server which operates in conjunctions with a front end and a data server for determining user access privileges.

46. The apparatus of claim 35, further comprising:

means for storing a user profile in a same graph as said data.

47. A user interface for use with a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said knowledge, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary by users, and connections between said knowledge and other units of knowledge, said user interface comprising:

a viewing tool for accessing said knowledge, wherein said viewing tool supports choosing topics that said user wants to learn about, viewing explanations provided to said user as a sequence of presentations, and annotating;

said viewing tool further comprising;

a module for allowing said user to navigate through links, see patterns in connections, and reorganize information according to multiple navigational schemes;

a module for allowing user to see the detailed local information, and also see how that information fits into a broader global context; and

a topic search engine for selecting one or more topics that a user wants to learn about.

48. The user interface of claim 47, further comprising:

a module for displaying a map of an area of topic space said user selects, showing what said user already knows and what is knowable;

wherein said map optionally illustrates a space of topics and subtopics as a two-dimensional landscape, with border, landmarks and links showing relationships between topics;

wherein said map optionally implements a coloring scheme showing said user'"'"'s prior knowledge and relative importance of a topic; and

wherein said interface is able to create a map specifically for said user.

49. The user interface of claim 47, further comprising:

a topic map, optionally drawn and colored in context, which depends both on what said user is trying to learn and on what other persons that said user trusts have judged to be important.

50. The user interface of claim 49, wherein said map allows said user to assess the size of each topic, and how long it takes said user to cover said topic;

wherein said map optionally shows paths that said user has traveled before and paths that other users have traveled before.

51. The user interface of claim 47, wherein said user interface allows said user to move through topic space by any of panning, zooming, and leaping from topic to related topic;

wherein said user can optionally zoom into relevant topics, look at their subtopics, and mark the things that are of interest, or that are already known.

52. The user interface of claim 47, wherein said user interface further comprises:

a simulation of a three-dimensional navigational space that said user can navigate through by any of moving forward/back, right/left, up/down, and rotating;

wherein said navigational space comprises a plurality of graphical objects which are any of three-dimensional and animated; and

which have sounds associated with them that said user begins to hear as he draws near said object;

wherein there are links between said objects representing relationships between concepts they represent;

wherein said links are optionally almost transparent;

wherein as said user moves near to an object, said links associated with said object optionally become more visible, fading as a chain of connections gets farther from said object;

wherein as the user approaches a link, links of that type optionally become more visible.

53. The user interface of claim 52, wherein objects are arranged in said navigational space in a systematic way according to a scheme;

wherein a vertical dimension optionally represents historical time; and

wherein a horizontal dimension optionally represents a theme;

wherein said organization scheme is not fixed such that when said user changes said organization scheme, said objects reorganize themselves in a new order.

54. The user interface of claim 52, wherein said user moves though said navigational space to find and examine objects of interest, to visualize their relationships, and to visualize a context into which they fit.

55. The user interface of claim 47, wherein said user interface uses a spatial metaphor at all levels of a topic tree;

wherein at a higher level a map represents topics maintained in a relationship to one another in a topic landscape, allowing said user to get to know an area of said topic landscape, and learn to navigate through said topic landscape;

wherein at a high level, said topic map changes slowly;

wherein at a lower, more detailed level, topics are more dynamic, and said topic map appears as a web of connections; and

wherein a same subtopic may appear on said map several times in different contexts.

56. The user interface of claim 47, wherein user visualization occur as any of a timeline, a map, a topic map, a hierarchical display, a display of personal bookmarks, and as a combination of said foregoing forms of visualization.

57. The user interface of claim 47, further comprising:

a display presented to said user;

means for allowing said user to select a search space;

wherein said search space is optionally any of a time line, a map view, an exploded map view, related topics, and a hierarchy.

58. The user interface of claim 47, wherein said user interface further comprises:

a simulation of a two-dimensional navigational space that said user can navigate through by any of moving right/left and up/down;

wherein said navigational space comprises a plurality of graphical objects which are any of two-dimensional and animated; and

which have sounds associated with them that said user begins to hear as he draws near said object;

wherein there are links between said objects representing relationships between concepts they represent;

wherein said links are optionally almost transparent;

wherein as said user moves near to an object, said links associated with said object optionally become more visible, fading as a chain of connections gets farther from said object;

wherein as the user approaches a link, links of that type optionally become more visible.

59. A method for use with a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said knowledge, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary by users, and connections between said knowledge and other units of knowledge, said method comprising the steps of:

when a data object is registered, using its type and content to generate a fast, unique hash value, which is used as an index into a registry;

wherein said hash value is used to identify and register a data object into said registry and is used as an index in said registry'"'"'s hash table.

60. A data object registry method for use with a knowledge base comprising knowledge, meta-knowledge that was created at a time of entry of said core content, and meta-knowledge in the form of one or more annotations that accumulate over time, said annotations including any of, but not limited to, usefulness of said knowledge, additional user opinions, certifications of veracity of said knowledge, commentary, and connections between said knowledge and other units of knowledge, said method comprising the steps of:

representing a plurality of registered data objects as a hash table entry;

wherein hash table entries identify a data object'"'"'s location, representation, and any associated metadata;

each hash table entry comprising an index hash, an optional cryptographically strong signature for verification and security, a data identifier, and a metadata identifier.

61. The method of claim 60, wherein said data identifier describes said data object'"'"'s type, length, and one or more representations of said object'"'"'s data.

62. The method of claim 60, wherein said hash table entry contains a metadata identifier which includes an indication of annotations of said data object.

63. The method of claim 60, wherein said index hash that is computed using padding algorithm applied to all data to ensure it is of sufficient length.

64. The method of claim 63, wherein said padding algorithm comprises the steps of:

receiving a request for N words of data;

performing a test to determine if there are N words of data in a buffer;

if there are N words of data in said buffer, returning said data;

if there are not N words of data in said buffer, filling as much of said buffer as possible with data;

performing a test to determine if said buffer is full;

if said buffer is full, returning said data;

if said buffer is not full, performing a test to determine if there are any data in said buffer;

if there are not any data in said buffer, returning a null value;

if there are data in said buffer, appending a byte value representing a number of words needed to fill said buffer and performing a test to determine if said buffer is full;

if said buffer is full, returning said data are returned;

if said buffer is not full, appending said data in said buffer up to a first added byte;

thereafter, performing a test to determine if said buffer is full;

if said buffer is full, returning said data;

if said buffer is not full, again appending a byte value representing a number of words needed to fill up said buffer; and

continuing.

65. The method of claim 60, further comprising the step of:

computing a signature of a data object using a data object type and content;

wherein said signature is computed using a cryptographically strong technique.

66. The method of claim 60, further comprising the step of:

providing a data identifier that contains a data object'"'"'s type, length, and representation;

wherein a data object'"'"'s representation contains at least one segment; and

wherein for each segment, said data identifier contains information denoting how to find a string of bits that represent a part of said data object.

67. The method of claim 60, wherein said data object is constructed by obtaining bits associated with each segment, concatenating them together sequentially, and interpreting them as specified by type;

wherein once all of said bits are collected, said bits are optionally verified by comparing an index hash computed from concatenated data and a type with said index hash stored in said hash table.

68. The method of claim 60, further comprising the step of:

providing a metadata identifier that contains at least one component that indicates type and location of one or more links annotating said data object;

wherein each metadata component can optionally specify multiple alternative locations where metadata are found; and

wherein each location optionally has a type specifying a format of metadata stored in that location.

69. The method of claim 68, wherein metadata for an object are constructed by obtaining data from one location indicated by each component;

wherein said metadata are then collected and interpreted based on each location'"'"'s type.