Query relaxation using external domain knowledge for query answering

Patent metadata

Field	Value
Publication number	US-11841867-B2
Application number	US-201916537496-A
Country	US
Kind code	B2
Filing date	Aug 9, 2019
Priority date	Aug 9, 2019
Publication date	Dec 12, 2023
Grant date	Dec 12, 2023

Abstract

A system, method, and computer readable medium perform a method of query relaxation. A query output from a conversational system is received. At least one search term in the query is identified. Instance data is output to the conversational system in response to determining that the instance data in a data store matches the at least one search term in the query. The computer device outputs the received query to an external domain-specific knowledge source in response to determining that the at least one search term does not the match instance data in the data store. The computer device receives the relaxed data matches from the external domain-specific knowledge source being semantically-related to at least one search term in the query based on a plurality of criteria associated with the query. The computing device generates a response to the query based on contextual information and structural information.

First claim

What is claimed is:

1. A non-transitory computer readable storage medium tangibly embodying a computer readable program code having computer readable instructions that, when executed, cause a computer device to execute a method of query relaxation, the method comprising:
receiving a query from a conversational system;
identifying at least one search term in the query;
outputting instance data to the conversational system in response to determining that the instance data in a data store matches the at least one search term in the query;
in response to determining that the at least one search term does not match the instance data in the data store, performing, by the computer device, a hybrid search of an external domain-specific knowledge source, the hybrid search comprising a radius approach, a similarity measure, and statistical information usage based on word frequency to find relaxed data matches that are semantically related to the at least one search term in the query; and
generating, by the computer device, a response to the query based on contextual information and structural information based on the relaxed data matches found from the hybrid search.

2. The non-transitory computer readable storage according to claim 1 , wherein the relaxed data matches are semantically-related terms received by the computer device based on structural information obtained from the external domain-specific knowledge source, the contextual information obtained from the query, and information content of the query.

3. The non-transitory computer readable storage medium according to claim 1 , wherein the external domain-specific knowledge source includes data in a form of at least one of taxonomies, ontologies, or semantic networks.

4. The non-transitory computer readable storage medium according to claim 1 , further comprising indexing data values stored in the data store with the instance data.

5. The non-transitory computer readable storage medium according to claim 1 , further comprising ingesting the external domain-specific knowledge source into a graphic database.

6. The non-transitory computer readable storage medium according to claim 1 , wherein the external domain-specific knowledge source is grouped in terms of concept, relation, and description.

7. The non-transitory computer readable storage medium according to claim 1 , wherein the instance data includes variations of semantically-related data obtained from the data store.

8. A system configured to perform query relaxation, comprising:
a plurality of data stores;
a translation index configured to:
receive a query output from a conversational system;
provide instance data indexing to data values stored in the plurality of data stores and to semantically related terms;
identify at least one search term in the query; and
output an instance data to the conversational system in response to determining that the data in the data stores matches the at least one search term in the query; and

an external domain-specific knowledge index configured to:
in response to determining that there is no match in the data stores of the at least one search term, performing a hybrid search comprising a radius approach, a similarity measure, and statistical information usage based on word frequency to find relaxed data matches that are semantically related to the at least one search term in the query; and
output the query to an external domain-specific knowledge source; and
receive and wherein the system generates a response to the query based on contextual information and structural information based on the relaxed data matches found from the hybrid search.

9. The system according to claim 8 , wherein the external domain-specific knowledge source includes data in a form of one or more of taxonomies, ontologies, or semantic networks.

10. The system according to claim 8 , further comprising a graphic database into which the external domain-specific knowledge source is ingested.

11. The system according to claim 8 , wherein the instance data includes variations of semantically-related data obtained from the data stores.

12. The system according to claim 8 , wherein the external domain-specific knowledge index is further configured to process relaxed data received from the external domain-specific knowledge source to determine one or more instance-concept frequencies in the external knowledge source and to determine one or more deep hierarchies in the external knowledge source.

13. The system according to claim 12 , wherein the external domain-specific knowledge source is instance-concept frequency populated.

14. The system according to claim 8 , wherein the translation index is further configured to parse entities and intents in the query received from the conversational system.

15. The system according to claim 8 , further wherein the configured to determine concepts, relationships, and descriptions of the external domain-specific knowledge source.

16. The system according to claim 15 , further configured to determine an hyponymy relation (isA) between classes and subclasses of the structural information, wherein the similarity is calculated according to:

IC
⁡
(
c
)

=

-
log

⁢

(

freq
⁡
(
c
)

)

;

⁢

freq
⁡
(
c
)

=

❘
"\[LeftBracketingBar]"

c

❘
"\[RightBracketingBar]"

+

∑

i
=
1

n

❘
"\[LeftBracketingBar]"

c
i

❘
"\[RightBracketingBar]"

;

⁢

sim
IC

(

c

Assignees

• IBM

Inventors

• Lei Chuan
• Ozcan Fatma
• Miller Dorian Boris
• Kreulen Jeffrey
• Geis Rebecca

Classifications

• G06F16/84

  Mapping; Conversion · CPC title

• G06F16/2468

  Fuzzy queries · CPC title

• G06F16/3329

  Natural language query formulation · CPC title

• G06F16/2465 Primary

  Query processing support for facilitating data mining operations in structured databases · CPC title

• G06F16/2272

  Management thereof · CPC title