Data graph interface

• US 9,798,829 B1
• Filed: 10/22/2013
• Issued: 10/24/2017
• Est. Priority Date: 10/22/2013
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Baisley, Donald Edward
• Primary Examiner(s): Hasan, Syed
• Assistant Examiner(s): Ahmed, Saba
• Application Number:
  US14/059,833
• Time in Patent Office: 1,463 Days
• Field of Search: 707805
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence

Abstract

Some implementations provide a user interface that allows non-technical users to import, create, and maintain a knowledge document. The knowledge document may include links to other knowledge documents or links to public data graphs. For example, a method may include generating data used to display tabular data to a user, a first column of the tabular data being associated with a first entity type and non-heading rows in the tabular data representing information for entities of the first entity type. The method may also include permitting the user to select a second column of the tabular data and provide a second entity type as an object type to be associated with a relationship represented by the second column. The method may include generating an entity in the knowledge document of the second entity type and linking an entity of the first entity type to the generated entity by the relationship.

Claims

1. A computer system comprising:

at least one processor; and

a memory storing;

a data graph including a plurality of first entities of a first entity type, wherein the first entity type has a first property of an alpha-numeric type, wherein at least some of the first entities have respective alpha-numeric values for the first property, and

instructions that, when executed by the at least one processor, cause the system to;

generate data used to display the data graph in a tabular format, where entities of the first entity type are displayed in a first column and the first property is displayed in a second column,

receive a selection of the second column,

receive, responsive to the selection of the second column, a new entity type for the first property from a user,

generate second entities having the new entity type, each second entity representing a unique alpha-numeric value for the first property from the respective alpha-numeric values, and

for a particular first entity of the plurality of first entities;

generate a labeled relationship in the data graph between the particular first entity and one of the second entities based on the alpha-numeric value of the first property for the particular first entity, the label being based on a heading of the second column, and

upon generating the relationship, remove the first property from the particular first entity.

2. The system of claim 1, wherein the first property has a fact template, the fact template being automatically generated based on semantic analysis of a column heading of tabular data and values in the column.

3. The system of claim 1, wherein the memory further stores instructions that, when executed by the at least one processor, cause the system to:

receive a subtype from the user for the first entity type and a natural language description for the subtype;

generate an entity in the data graph that represents the subtype;

parse the natural language description to determine criteria for the subtype; and

determine that first entities meeting the criteria are members of an extension of the subtype.

4. The system of claim 3, wherein the criteria specifies a property, a relationship and an object entity, or at least two properties and a logical operator.

5. The system of claim 1, wherein the memory further stores instructions that, when executed by the at least one processor, cause the system to:

receive a subtype for the first entity type;

generate an entity that represents the subtype;

receive an indication that a particular first entity is an instance of the subtype; and

generate a relationship in the data graph between the particular first entity and the entity that represents the subtype.

6. The system of claim 1, wherein the memory further stores instructions that, when executed by the at least one processor, cause the system to:

receive an existing entity type as the type for a second property;

determine that a certain first entity has an alpha-numeric value for the second property that corresponds to an entity of the existing entity type;

generate a relationship in the data graph between the certain first entity and the entity of the existing entity type, the relationship representing the second property; and

remove the second property from the certain first entity.

7. The system of claim 1, further comprising a user interface that generates information used to display the first entities, the information for each first entity including:

alpha-numeric values for the first entity for each associated property; and

entities related to the first entity.

8. The system of claim 7, wherein the entities related to the first entity appear as links that enable the user of the user interface to initiate display of information for the related entities.

9. The system of claim 7, wherein the user interface further generates information used to display the entity types in the data graph and to select one of the entity types as the first entity type.

10. The system of claim 1, the memory further stores instructions that, when executed by the at least one processor, cause the system to:

receive a second entity type for a second property, the second entity type being an entity type from an external data graph; and

for each of the plurality of first entities with a second alpha-numeric value for the second property;

match the second alpha-numeric value to an entity in the external data graph that has the second entity type, and

generate a relationship in the data graph between the first entity and the entity in the external data graph.

11. The system of claim 10, further comprising a user interface that generates information used to display the first entities, the information for each first entity including:

alpha-numeric values for properties associated with the first entity; and

entities related to the first entity, including the entities in the external data graph.

12. The system of claim 10, further comprising a user interface that adds properties in the data graph for the entities in the external data graph.

13. A method comprising:

generating, using at least one processor, data used to display tabular data to a user, a first column of the tabular data being associated with a first entity type and non-heading rows in the tabular data representing entities of the first entity type in a knowledge document;

permitting the user to select a second column of the tabular data and provide a second entity type as an object type to be associated with a labeled relationship represented by a heading of the second column;

generating, using the at least one processor, an entity in the knowledge document of the second entity type that represents a value in a non-heading row in the tabular data;

linking an entity of the first entity type that is associated with a non-heading row to the generated entity by the relationship; and

storing an updated knowledge document in a memory.

14. The method of claim 13, further comprising:

modifying a display of the tabular data responsive to storing the updated knowledge document so that values in the second column are selectable.

15. The method of claim 13, further comprising:

permitting the user to select a third column of the tabular data and provide the entity type as an object type to be associated with a relationship represented by the third column;

for each non-heading row;

determining that the third column for the non-heading row has a value,

identifying a second entity of the first entity type that corresponds to the value,

and linking a first entity of the first entity type that is associated with the non-heading row to the second entity using the relationship represented by the third column; and

storing updates to the knowledge document in the memory.

16. The method of claim 13, further comprising:

receiving a natural language computation for a subtype, the subtype being associated with the first entity type;

parsing the natural language computation to determine criteria for the subtype;

determining a plurality of entities of the first entity type meet the criteria; and

displaying the plurality of entities in a tabular display for the subtype.

17. The method of claim 13, wherein the second entity type does not exist prior to the user selecting the second column.

18. The method of claim 13, further comprising:

receiving an identification of the tabular data from the user;

defining the first entity type based on an initially occurring column of data in the tabular data;

generating the entities of the first entity type based on row values in the initially occurring column of data;

generating properties associated with the entities of the first entity type based on row values in remaining columns of data; and

storing the entities and their associated properties as the knowledge document.

19. The method of claim 13, further comprising:

permitting the user to add a new column to the tabular data by supplying a name and a natural language computation, the natural language computation identifying a particular relationship;

parsing the computation to determine the particular relationship;

determining that the computation specifies an inverse relationship;

determining a plurality of entities that satisfy the inverse relationship for a second entity of the first entity type; and

displaying representations for the plurality of entities as values for the new column in a row of the tabular display associated with the second entity of the first entity type.

20. The method of claim 13, further comprising:

permitting the user to add a new column to the tabular data by supplying a name and identifying a third entity type from a data graph distinct from the knowledge document; and

permitting the user to associate an entity from the data graph with an entity of the first entity type using a relationship represented by the new column.

21. A computer-implemented method comprising:

storing a first data graph in memory of a computing device, the first data graph including a plurality of first entities of a first entity type, wherein the first entity type has one or more associated properties, a first property of the one or more properties having a text type, wherein at least some of the first entities have respective values for the first property;

receiving an entity type for the first property;

determining that the received entity type is from a second data graph; and

for each of the plurality of first entities with a value for the first property;

matching the value to an entity in the second data graph that has the received entity type, and

generating a relationship in the first data graph between the first entity and the entity in the second data graph.

22. The method of claim 21, further comprising:

receiving a new property for association with the entity type from the second data graph;

receiving a value for the new property associated for a particular entity in the second data graph; and

storing the value for the new property as part of the first data graph, wherein the new property and the value for the new property are unknown to second data graph.

23. The method of claim 22, further comprising:

generating data used to display property values associated with the entity from the second data graph and the value for the new property.