Method and system to enable inferencing for natural language queries of configuration management databases
View Patent ↗Disclosed are embodiments of systems and methods to derive a semantic network from a CMDB relationship graph which can then be queried in a natural way from a linguistic standpoint (i.e., using natural language queries). Because disclosed embodiments combine natural language queries with an inferencing engine the disclosed systems and methods automatically “connect the dots” between disparate pieces of information and can allow for a richer user experience. In general, CMDB graph relationships can be converted into semantic networks. Once a semantic network is created, queries can be phrased to leverage the inferential relationships between objects in the semantic network.
1. A method of performing a natural language query attains a database using a computer system comprising one or more programmable processing units on one or more devices communicatively coupled to each other, the method comprising:
receiving a natural language query from an end user at the computer system;
parsing the natural language query to identify one or more keywords;
generating a structured query based at least in part on the one or more keywords;
obtaining at least a portion of a semantic network representation comprising one or more relationships associating different items of data in the database;
generating one or more inference rules by normalizing the one or more relationships against a language elements dictionary;
initiating the structured query thereby generating a first results set; and
applying the one or more inference rules to the first results set to produce a second results set.
2. The method of claim 1 wherein obtaining comprises retrieving from an already populated semantic network.
3. The method of claim 1 wherein obtaining comprises generating a semantic network from information available in the database.
4. The method of claim 1 wherein the database is a Configuration Management Database (CMDB).
5. The method of claim 4 wherein the CMDB is a federated CMDB.
6. The method of claim 1 wherein applying the one or more inference rules comprises applying inference rules using a breadth-wise traversal technique.
7. The method of claim 1 wherein applying the one or more inference rules comprises applying inference rules using a depth-wise traversal technique.
8. The method of claim 7 wherein the one or more inference rules are iteratively applied until a branch of the search tree is exhausted by reaching a leaf node or reaching a relationship branch not of interest.
9. The method of claim 1 further comprising processing the second result set to translate the data of the second results set into a natural language format prior to presentation to an end user.
10. The method of claim 1 further comprising:
determining whether the structured query can be generated from the received natural language query; and
initiating a prompt to a user to refine the natural language query when a structured query cannot be properly generated.
11. The method of claim 1 where initiating the structured query and receiving a first results set are performed substantially in parallel with the acts of obtaining at least a portion of a semantic network and generating one or more inference rules.
12. A computer system comprising one or more programmable processing units on one or more devices communicatively coupled to each other wherein the one or more programmable processing units are programmed to:
receive a natural language query from an end user at the computer system;
parse the natural language query to identify one or more keywords;
generate a structured query based at least in part on the one or more keywords;
obtain at least a portion of a semantic network representation comprising one or more relationships associating different items of data in the database;
generate one or more inference rules by normalizing the one or more relationships against a language elements dictionary;
initiate the structured query thereby generating a first results set; and
apply the one or more inference rules to the first results set to produce a second results set.
13. The computer system of claim 12 wherein the programmed act of obtaining comprises retrieving from an already populated semantic network.
14. The computer system of claim 12 wherein programmed act of obtaining comprises generating a semantic network from information available in the database.
15. The computer system of claim 12 wherein the database is a Configuration Management Database (CMDB).
16. The computer system of claim 15 wherein the CMDB is a federated CMDB.
17. The computer system of claim 12 wherein the programmed act of applying the one or more inference rules comprises applying inference rules using a breadth-wise traversal technique.
18. The computer system of claim 12 wherein the programmed act of applying the one or more inference rules comprises applying inference rules using a depth-wise traversal technique.
19. The computer system of claim 18 wherein the one or more inference rules are iteratively applied until a branch of the search tree is exhausted by reaching a leaf node or reaching a relationship branch not of interest.
20. The computer system of claim 12 further comprising a processing unit programmed to process the second result set to translate the data of the second results set into a natural language format prior to presentation to an end user.
21. The computer system of claim 12 further comprising a processing unit programmed to:
determine whether the structured query can be generated from the received natural language query; and
initiate a prompt to a user to refine the natural language query when a structured query cannot be properly generated.
22. The computer system of claim 12 where the programmed acts of initiating the structured query and receiving a first results set are performed substantially in parallel with the programmed acts of obtaining at least a portion of a semantic network and generating one or more inference rules.
23. A non-transitory computer readable medium having computer-executable instructions stored thereon for instructing a computer system to perform the method of claim 1 .
24. A computer network comprising:
a plurality of processing units communicatively coupled to a computer network; and
a first processing unit configured to perform at least a portion of the method of claim 1 wherein the entire method of claim 1 is performed collectively by the plurality of processing units.