KNOWLEDGE REPRESENTATION ON ACTION GRAPH DATABASE

• US 20160292304A1
• Filed: 09/29/2015
• Published: 10/06/2016
• Est. Priority Date: 04/01/2015
• Status: Active Grant

• Current Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Original Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Inventors: KARTHA, Rajesh, V, Supriya, CHACKO, Viju, SARKAR, Shampa
• Granted Patent: US 10,019,538 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 5/027 Frames

Abstract

Knowledge representation in multi-layered database includes systems and methods for storing and retrieving data in the multi-layered database. In the multi-layered database, an action graph database includes participant-entity nodes corresponding to real world entities and action nodes corresponding to action capabilities of the real world entities. Each of the participant-entity nodes and the action nodes is associated with properties, relationships, and relationship properties. Underlying the action graph layer is a standard graph layer that stores nodes, node properties associated with the nodes, edges, and edge properties associated with the edges, wherein the nodes correspond to the participant-entity nodes and the action nodes. Further, underlying the standard graph layer is a backend database layer that stores corresponding data and metadata.

Claims

1. A system for storing data using a knowledge representation scheme in a multi-layered database, the system comprising:

a processor;

an operation module coupled to the processor to;

receive attributes, from a user, characterizing properties and relationships of a real world entity and at least of an action capability of the real world entity for storage in the multi-layered database;

store a participant-entity node corresponding to the real world entity in an action graph layer of the multi-layered database, wherein, based on the attributes of the real world entity, the participant-entity node is associated with participant properties that characterize the participant-entity node, and participant relationship properties that relates the participant-entity node to at least one other participant-entity node;

store an action node corresponding to the action capability of the real world entity in the action graph layer of the multi-layered database, wherein, based on the attributes of the action capability, the action node is associated with action properties that characterize the action node, action relationship that relates the action node to at least one other action node, action relationship properties that characterize the action relationship, action-participant relationship that relates the action node to the participant-entity node, and action-participant relationship properties that characterize the action-participant relationship;

a decomposer module coupled to the processor to;

decompose the participant-entity node, the participant-properties, the action node, and the action-properties of the action graph layer into nodes and properties associated with the corresponding nodes;

decompose the participant-relationship, the action-relationship,

the action-participant relationship, and corresponding relationship properties of the relationships in the action graph layer, into edges connecting the nodes and edge properties associated with the corresponding edges respectively;

store the nodes, the edges, the node properties, and the edge properties in a standard graph layer of the multi-layered database;

decompose further the nodes, the edges, the node properties, and the edge properties of the standard graph layer into data and metadata associated with the data; and

store the data and the metadata in a backend database layer of the multi-layered database.

2. The system as claimed in claim 1, wherein the participant-entity node is at least one of a subject node, an object node, and a primitive literal string node, wherein the action-participant relationships comprise at least one of a modal relationship and a general relationship that connect at least an action node to one or more participant-entity nodes to facilitate many-to-many participant-entity mapping, wherein the modal relationship relates the action node to the subject node, and wherein the general relationship relates the action node to at least an object node or a primitive literal string node.

3. The system as claimed in claim 1, wherein the action relationship comprise at least one of an inheritance relationship, step relationship, constituent (synonym) relationship, and a like relationship, wherein the inheritance relationship facilitates a child action node to inherit properties and relationships of its parent action node, wherein the step relationship facilitates an action node to be related to another action node comprising a particular action step, wherein the constituent relationship facilitates an action node to be related to one or more action nodes representing synonyms of action associated with that action node, and wherein the like relationship facilitates an action node to be related to at least another action node with similar properties.

4. The system as claimed in claim 1, wherein the action properties further comprises parameters comprising at least a pre-condition parameter, a step parameter, and a quantity parameter, wherein the precondition parameter captures pre-conditional actions necessary for an action to be performed, a step parameter stores the level of granularity up to which the user would like to retrieve information in case of read operations, and the quantity parameter quantizes the action node in terms of its participant-entity quantities.

5. The system as claimed in claim 1, wherein the system is further coupled to an indexed seek module to index the nodes and edges of the standard graph layer wherein the nodes being decomposed from the participant-entity nodes, and the action nodes of the action graph layer, and wherein the edges being decomposed from the action relationship, participant relationship, and action-participant relationship of the action graph layer, based on at least one indexing technique, wherein the at least one indexing technique is pluggable to the backend database layer.

6. A system for retrieving capability information by a user using a knowledge representation scheme in a multi-layered database, the system comprising:

a processor;

an application programming interface (API) module coupled to the processor to receive a query from the user to retrieve the capability information from the multi-layered database, wherein the query comprises at least a participant-entity node corresponding to a real world entity and at least an action node corresponding to an action capability of the real world entity; and

a traversal module coupled to the processor to identify a traversing logic, based on the query, one or more participant-entity nodes and one or more action nodes in an action graph layer of the multi-layered database, wherein the traversing logic corresponds to the one or more participant-entity nodes associated with participant properties and the action nodes associated with action properties, and wherein a participant relationship relates the participant entity node to the at least one other participant-entity node, the participant properties characterize a participant-entity node, and wherein an action relationship relates the action node to the at least one other action node, an action relationship properties that characterize the action relationship, action-participant relationship that relates the action node to the at least one participant-entity node, and action-participant relationship properties that characterize the action-participant relationship, and wherein the action properties characterize the action node, further comprising at least one of a pre-condition parameter, a step parameter and a quantity parameter;

a seek module coupled to the processor to perform a search of capability information associated with the query in the action graph layer, based on at least the action node, the participant-entity node and the traversing logic;

an authorization module coupled to the processor to determine access privileges for the user to access the capability information, wherein the access privileges are based on user role; and

a retrieval module coupled to the processor to retrieve the capability information for providing to the user when the user is privileged to access the capability information.

7. The system as claimed in claim 6, wherein the retrieval module is further configured to compose the capability information, wherein the composing the capability information comprises:

identifying, in the action graph layer, the traversing logic to search the participant-entity node and the action node comprised in the query;

identifying, in the standard graph layer, one or more nodes corresponding to the participant-entity node and the action node, edges connecting the one or more nodes, and the properties associated with the one or more nodes and the edges;

obtaining data and metadata corresponding to the one or more nodes and the edges from the backend database layer; and

compiling the one or more nodes based on the data and the metadata, and the one or more participant-entity nodes and the one or more action nodes based on the one or more nodes to form the capability information.

8. The system as claimed in claim 6, wherein the seek module searches the capability information by performing a search based on the query in an index of the standard graph layer, and wherein the index is a segmented index implementing a plurality of indexing techniques.

9. The system as claimed in claim 6, wherein the seek module retrieves the capability information by performing a non-indexed search in the standard graph layer

10. A computer implemented method for storing data using a knowledge representation scheme in a multi-layered database, the method comprising:

receiving, by a processor, attributes of a real world entity and at least an action capability of the real world entity for storage in the multi-layered database, wherein the attributes characterize properties and relationships of the real world entity and the properties and the relationships of the action capability of the real world entity;

storing, by the processor, a participant-entity node corresponding to the real world entity in an action graph layer of the multi-layered database, wherein, based on the attributes of the real world entity, the participant-entity node is associated with participant properties that characterize the participant-entity node, and participant relationship properties that relates the participant-entity node to at least one other participant-entity node;

storing, by the processor, an action node corresponding to the action capability of the real world entity in the action graph layer of the multi-layered database, wherein, based on the attributes of the action capability, the action node is associated with action properties that characterize the action node, action relationship that relates the action node to at least one other action node, action relationship properties that characterize the action relationship, action-participant relationship that relates an action node to the participant-entity node, and the action-participant relationship properties that characterize the action-participant relationship;

decomposing, by the processor, the participant-entity node, the participant-properties, the action node, and the action-properties of the action graph layer into nodes and properties associated with the corresponding nodes;

decomposing, by the processor, the participant-relationship, the action-relationship, and the action-participant relationship and the corresponding relationship properties of the relationships in the action graph layer, into edges connecting the nodes and edge properties associated with the corresponding edges respectively;

storing, by the processor, the nodes, the edges, the node properties, and the edge properties in a standard graph layer of the multi-layered database;

decomposing further, by the processor, the nodes, the edges, the node properties, and the edge properties into data and metadata associated with the data, and

storing, by the processor, the data and the metadata in a backend database layer of the multi-layered database.

11. The method as claimed in claim 10, wherein the participant-entity node is at least one of a subject node, an object node, and a primitive literal string node, and wherein the action-participant relationships comprise at least one of a modal relationship and a general relationship that connect at least an action node to the one or more participant-entity nodes, for facilitating many-to-many participant-entity mapping, wherein the modal relationship relates the action node to the subject node, and wherein the general relationship may include a general association between at least an action node and an object node or a primitive literal string node.

12. The method as claimed in claim 10, wherein the action relationship comprise at least one of an inheritance relationship, step relationship, constituent (synonym) relationship, and a like relationship, wherein the inheritance relationship facilitates a child action node to inherit properties and relationships of its parent action node, wherein the step relationship facilitates an action node to be related to another action node comprising a particular action step, wherein the constituent relationship facilitates an action node to be related to the one or more action nodes representing synonyms of action associated with that action node, and wherein the like relationship facilitates an action node to be related to at least another action node with similar properties.

13. The method as claimed in claim 10, comprising updating, by the processor, the action properties comprising parameters such as at least a pre-condition parameter, a step parameter and a quantity parameter, wherein a precondition parameter captures pre-conditional actions necessary for an action to be performed, a step parameter stores the level of granularity up to which the user would like to retrieve information in case of read operations, and a quantity parameter quantizes the action node in terms of its participant-entity quantities.

14. The method as claimed in claim 10 comprising indexing the nodes and edges of the standard graph layer wherein the nodes being decomposed from the participant-entity nodes, and the action nodes of the action graph layer, and wherein the edges being decomposed from the action relationship, the participant relationship and the action-participant relationship of the action graph layer, based on at least one indexing technique, wherein the at least one indexing technique is pluggable to the backend database layer.

15. A computer implemented method for retrieving capability information by a user using a knowledge representation scheme in a multi-layered database, the method comprising:

receiving, by a processor, a query from the user to retrieve the capability information from the multi-layered database, wherein the query comprises at least a participant-entity node corresponding to a real world entity and at least an action node corresponding to an action capability of the real world entity;

based on the query, classifying, by the processor, the query in one of a comparison query, an inference query, and a direct query;

based on the classified query, identifying, by the processor, a traversing logic to traverse an action graph layer of the multi-layered database, wherein the traversing logic corresponds to the one or more participant-entity nodes associated with participant properties and the action nodes associated with action properties, and wherein a participant relationship relates the participant entity node to the at least one other participant-entity node, the participant properties characterize the participant-entity node, and wherein an action relationship relates the action node to at least one other action node, an action relationship properties that characterize the action relationship, action-participant relationship that relates the action node to the at least one participant-entity node, and action-participant relationship properties that characterize the action-participant relationship, and wherein the action properties characterize the action node, further comprising at least one of a pre-condition parameter, a step parameter and a quantity parameter;

searching, by the processor, the capability information associated with the classified query in the action graph layer of the multi-layered database, based on the action node, the participant-entity node and the traversing logic;

determining, by the processor, access privileges for the user to access the capability information, wherein the access privileges are based on user role; and

retrieving, by the processor, the capability information for providing to the user when the user is privileged to access the capability information.

16. The method as claimed in claim 15 comprising composing, by the processor, the capability information, by:

identifying, by the processor, in the action graph layer, the traversing logic to search the participant-entity node and the action node comprised in the query;

identifying, by the processor, in the standard graph layer, one or more nodes corresponding to the participant-entity node and the action node, edges connecting the one or more nodes, and properties associated with the one or more nodes;

obtaining, by the processor, the data and metadata corresponding to the one or more nodes and the edges from the backend database layer; and

compiling, by the processor, the one or more nodes based on the data and the metadata, and the one or more participant-entity nodes and the one or more action nodes based on the one or more nodes to form the capability information.

17. The method as claimed in claim 15, wherein searching the capability information includes performing, by the processor, a search based on the query in an index of the standard graph layer, wherein the index is a segmented index implementing a plurality of indexing techniques.

18. The method as claimed in claim 15, wherein searching the capability information includes performing, by the processor, a non-indexed search in the standard graph layer.

19. A non-transitory computer readable medium comprising instructions executable by a processor to perform a method for storing data using a knowledge representation scheme in a multi-layered database, the method comprising:

receiving, by a processor, attributes of a real world entity and at least an action capability of the real world entity for storage in the multi-layered database, wherein the attributes characterize the properties and the relationships of the real world entity and the properties and the relationships of the action capability of the real world entity;

storing, by the processor, a participant-entity node corresponding to the real world entity in an action graph layer of the multi-layered database, wherein, based on the attributes of the real world entity, the participant-entity node is associated with participant-properties that characterize the participant-entity node, and participant relationship properties that relates the participant-entity node to at least one other participant-entity node;

storing, by the processor, an action node corresponding to the action capability of the real world entity in the action graph layer of the multi-layered database, wherein, based on the attributes of the action capability, the action node is associated with action properties that characterize the action node, action relationship that relates the action node to at least one other action node, action relationship properties that characterize the action relationship, action-participant relationship that relates an action node to the participant-entity node and an action-participant relationship properties that characterize the action-participant relationship;

decomposing, by the processor, the participant-entity node, the participant-properties, the action node, and the action-properties of the action graph layer into nodes and properties associated with the corresponding nodes;

decomposing, by the processor, the participant-relationship, the action-relationship, and the action-participant relationship and the corresponding relationship properties of the relationships in the action graph layer, into edges connecting the nodes and edge properties associated with the corresponding edges respectively;

storing, by the processor, the nodes, the edges, the node properties, and the edge properties in a standard graph layer of the multi-layered database;

decomposing further, by the processor, the nodes, the edges, the node properties, and the edge properties into the data and metadata associated with the data; and

storing, by the processor, the data and the metadata in a backend database layer of the multi-layered database.

20. A non-transitory computer readable medium comprising instructions executable by the processor to perform a method for retrieving capability information using a knowledge representation scheme in a multi-layered database, the method comprising

receiving, by a processor, a query from a user to retrieve the capability information from the multi-layered database, wherein the query comprises at least a participant-entity node corresponding to a real world entity and at least an action node corresponding to an action capability of the real world entity;

based on the query, classifying, by the processor, the query in one of a comparison query, an inference query, and a direct query;

based on the classified query, identifying, by the processor, a traversing logic to traverse an action graph layer of the multi-layered database, wherein the traversing logic corresponds to the one or more participant-entity nodes associated with participant properties and the action nodes associated with action properties, and wherein a participant relationship relates the participant entity node to the at least one other participant-entity node, the participant properties characterize a participant-entity node, and wherein an action relationship relates the action node to the at least one other action node, an action relationship properties that characterize the action relationship, action-participant relationship that relates the action node to the at least one participant-entity node, and action-participant relationship properties that characterize the action-participant relationship, and wherein the action properties characterize an action node, further comprising at least one of a pre-condition parameter, a step parameter and a quantity parameter;

searching, by the processor, the capability information associated with the classified query in the action graph layer of the multi-layered database, based on the action node, the participant-entity node, and the traversing logic;

determining, by the processor, access privileges for the user to access the capability information, wherein the access privileges are based on user role; and

retrieving, by the processor, the capability information for providing to the user when the user is privileged to access the capability information.