Granted Patent B2

US 12,748,790 · App. 17/940,494 · Granted Sep 29, 2026

Inventors: Karina Elayne Kervin (Sacramento, CA); Manish Kesarwani (Bengaluru, IN); Satyajeet Raje (White Plains, NY); Nergal Issaie (San Jose, CA)

Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION

G06F16/3344G06F16/3328G06F16/3329

Abstract

Claims (67)

1 . A computer-implemented method comprising:

generating an intent string representative of a user's intention based on an input string received as an input by the user, wherein the input includes a query request directed to querying a graph database;

generating a search string that includes text from the input string and text from the intent string, wherein the generating the search string further comprises:

processing the input string using a natural language processing technique to extract a phrase from the input string, producing an extracted phrase;

generating a vector embedding of the extracted phrase;

extracting one or more phrases from metadata associated with the graph database;

computing one or more embeddings for the one or more phrases extracted from the metadata;

identifying a closest phrase based on a similarity comparison between the vector embedding of the extracted phrase and the one or more embeddings for the one or more phrases extracted from the metadata; and

appending text from the closest phrase to the search string;

executing a text-based query against the graph database using the search string;

executing an ontology-based query against the graph database using the search string;

generating a combined set of search results from a first set of search results and a second set of search results, wherein the first set of search results is from the text-based query, and wherein the second set of search results is from the ontology-based query;

generating final relevance scores for the combined set of search results by adjusting at least a portion of preliminary relevance scores of the combined set of search results; and

generating a query response that includes the combined set of search results ranked according to the final relevance scores.

2 . The computer-implemented method of claim 1 , wherein the generating of the intent string comprises processing the input string using a natural language processing technique.

3 . The computer-implemented method of claim 1 , wherein the generating of the intent string comprises extracting a phrase from the input string and generating a query embedding representative of the phrase.

4 . The computer-implemented method of claim 3 , wherein the generating of the intent string comprises extracting metadata from the graph database and generating a data embedding representative of the metadata.

5 . The computer-implemented method of claim 4 , wherein the generating of the intent string comprises ranking the data embedding among other data embeddings based on a comparison of the data embedding to the query embedding.

6 . The computer-implemented method of claim 1 , wherein the text-based query comprises matching the text from the input string and text from the intent string to textual properties of indexed nodes and relationships of the graph database.

7 . The computer-implemented method of claim 1 , wherein the ontology-based query comprises generating a network template that includes text from the input string and text from the intent string and matching nodes and relationships of the graph database to the network template.

8 . The computer-implemented method of claim 1 , wherein the adjusting of at least a portion of the preliminary relevance scores comprises normalizing the preliminary relevance scores of the first and second sets of search results.

9 . The computer-implemented method of claim 1 , further comprising:

identifying a table that is included in both the first set of search results and the second set of search results,

wherein the adjusting of at least a portion of the preliminary relevance scores comprises increasing the preliminary relevance score of the table responsive to identifying that the table is included in both the first set of search results and the second set of search results.

10 . The computer-implemented method of claim 9 , wherein the adjusting of the preliminary relevance score of the table comprises increasing the preliminary relevance score by a predetermined percentage of the preliminary relevance score.

11 . A computer program product comprising one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by a processor to cause the processor to perform operations comprising:

generating an intent string representative of a user's intention based on an input string received as an input by the user, wherein the input includes a query request directed to querying a graph database;

generating a search string that includes text from the input string and text from the intent string, wherein the generating the search string further comprises:

processing the input string using a natural language processing technique to extract a phrase from the input string, producing an extracted phrase;

generating a vector embedding of the extracted phrase;

extracting one or more phrases from metadata associated with the graph database;

computing one or more embeddings for the one or more phrases extracted from the metadata;

identifying a closest phrase based on a similarity comparison between the vector embedding of the extracted phrase and the one or more embeddings for the one or more phrases extracted from the metadata; and

appending text from the closest phrase to the search string;

executing a text-based query against the graph database using the search string;

executing an ontology-based query against the graph database using the search string;

generating a combined set of search results from a first set of search results and a second set of search results, wherein the first set of search results is from the text-based query, and wherein the second set of search results is from the ontology-based query;

generating final relevance scores for the combined set of search results by adjusting at least a portion of preliminary relevance scores of the combined set of search results; and

generating a query response that includes the combined set of search results ranked according to the final relevance scores.

12 . The computer program product of claim 11 , wherein the program instructions are stored in a computer readable storage device in a data processing system, and wherein the program instructions are transferred over a network from a remote data processing system.

13 . The computer program product of claim 11 , wherein the program instructions are stored in a computer readable storage device in a server data processing system, and wherein the program instructions are downloaded in response to a request over a network to a remote data processing system for use in a computer readable storage device associated with the remote data processing system, further comprising:

program instructions to meter use of the program instructions associated with the request; and

program instructions to generate an invoice based on the metered use.

14 . The computer program product of claim 11 , wherein the generating of the intent string comprises processing the input string using a natural language processing technique.

15 . The computer program product of claim 11 , wherein the generating of the intent string comprises extracting a phrase from the input string and generating a query embedding representative of the phrase.

16 . The computer program product of claim 11 , further comprising:

identifying a table that is included in both the first set of search results and the second set of search results,

wherein the adjusting of at least a portion of the preliminary relevance scores comprises increasing the preliminary relevance score of the table responsive to identifying that the table is included in both the first set of search results and the second set of search results.

17 . A computer system comprising a processor and one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by the processor to cause the processor to perform operations comprising:

generating an intent string representative of a user's intention based on an input string received as an input by the user, wherein the input includes a query request directed to querying a graph database;

generating a search string that includes text from the input string and text from the intent string, wherein the generating the search string further comprises:

processing the input string using a natural language processing technique to extract a phrase from the input string, producing an extracted phrase;

generating a vector embedding of the extracted phrase;

extracting one or more phrases from metadata associated with the graph database;

computing one or more embeddings for the one or more phrases extracted from the metadata;

identifying a closest phrase based on a similarity comparison between the vector embedding of the extracted phrase and the one or more embeddings for the one or more phrases extracted from the metadata; and

appending text from the closest phrase to the search string;

executing a text-based query against the graph database using the search string;

executing an ontology-based query against the graph database using the search string;

generating a combined set of search results from a first set of search results and a second set of search results, wherein the first set of search results is from the text-based query, and wherein the second set of search results is from the ontology-based query;

generating final relevance scores for the combined set of search results by adjusting at least a portion of preliminary relevance scores of the combined set of search results; and

generating a query response that includes the combined set of search results ranked according to the final relevance scores.

18 . The computer system of claim 17 , wherein the generating of the intent string comprises processing the input string using a natural language processing technique.

19 . The computer system of claim 17 , wherein the generating of the intent string comprises extracting a phrase from the input string and generating a query embedding representative of the phrase.

20 . The computer system of claim 17 , further comprising:

identifying a table that is included in both the first set of search results and the second set of search results,

wherein the adjusting of at least a portion of the preliminary relevance scores comprises increasing the preliminary relevance score of the table responsive to identifying that the table is included in both the first set of search results and the second set of search results.

Assignments (1)

Continuity (1)

References Cited (39)