Semantic text analysis for glossary maintenance
View Patent ↗A computer-implemented method (CIM), according to one embodiment, includes causing a first search to be performed on a first knowledge base for a first descriptive name, and extracting sentences from results of the first search. The method further includes running at least one predetermined deep learning model on the results of the first search for determining similarity scores for each of the extracted sentences, where each of the similarity scores defines a similarity score for an associated one of the extracted sentences and a first glossary. A first of the determined similarity scores is used to map terms of a second glossary to terms of the first glossary to enhance context provided in search results generated using the glossaries.
1 . A computer-implemented method (CIM), the CIM comprising:
causing a first search to be performed on a first knowledge base for a first descriptive name;
extracting sentences from results of the first search,
wherein the extracted sentences include labels associated with the first descriptive name and descriptions of the labels;
running at least one predetermined deep learning model on the results of the first search for determining similarity scores for the extracted sentences, wherein each of the similarity scores defines a similarity score for an associated one of the extracted sentences and a first glossary;
determining a first of the determined similarity scores, wherein the first determined similarity score has a relatively greater similarity score than the other determined similarity scores; and
using the first determined similarity score to map terms of a second glossary to terms of the first glossary to enhance context provided in search results generated using the glossaries.
2 . The CIM of claim 1 , wherein using the first determined similarity score to map terms of the second glossary to terms of the first glossary includes: mapping the label of the extracted sentence associated with the first determined similarity score to at least one of the terms of the second glossary, and comprising: in response to receiving, from a first device, a request for a second search to be performed on the at least one of the terms of the second glossary, causing the label of the extracted sentence associated with the first determined similarity score to be returned to the first device.
3 . The CIM of claim 2 , wherein the first glossary is a financial industry business ontology.
4 . The CIM of claim 1 , wherein the predetermined deep learning model run on the results of the first search is a Prompt-based Contrastive Learning for Sentence Embeddings (PromCSE) model.
5 . The CIM of claim 4 , wherein a second predetermined deep learning model is run on the results of the first search for determining the similarity scores for the extracted sentences, wherein the second predetermined deep learning model is a Term Frequency-Inverse Document Frequency (TF-IDF) model.
6 . The CIM of claim 1 , wherein the first descriptive name is selected from the group consisting of: a column name in a database, a header in a comma-separated values (CSV) formatted text file, and a categorical variable in a CSV formatted text file.
7 . The CIM of claim 1 , wherein the first knowledge base is an open and editable knowledge base accessible on the Internet.
8 . A computer program product (CPP), the CPP comprising:
a set of one or more computer-readable storage media; and
program instructions, collectively stored in the set of one or more computer-readable storage media, for causing a processor set to perform the following computer operations:
cause a first search to be performed on a first knowledge base for a first descriptive name;
extract sentences from results of the first search,
wherein the extracted sentences include labels associated with the first descriptive name and descriptions of the labels;
run at least one predetermined deep learning model on the results of the first search for determining similarity scores for the extracted sentences, wherein each of the similarity scores defines a similarity score for an associated one of the extracted sentences and a first glossary;
determine a first of the determined similarity scores, wherein the first determined similarity score has a relatively greater similarity score than the other determined similarity scores; and
use the first determined similarity score to map terms of a second glossary to terms of the first glossary to enhance context provided in search results generated using the glossaries.
9 . The CPP of claim 8 , wherein using the first determined similarity score to map terms of the second glossary to terms of the first glossary includes: mapping the label of the extracted sentence associated with the first determined similarity score to at least one of the terms of the second glossary, and the CPP comprising: program instructions, collectively stored in the set of one or more storage media, for causing the processor set to perform the following computer operations: in response to receiving, from a first device, a request for a second search to be performed on the at least one of the terms of the second glossary, cause the label of the extracted sentence associated with the first determined similarity score to be returned to the first device.
10 . The CPP of claim 9 , wherein the first glossary is a financial industry business ontology.
11 . The CPP of claim 8 , wherein the predetermined deep learning model run on the results of the first search is a Prompt-based Contrastive Learning for Sentence Embeddings (PromCSE) model.
12 . The CPP of claim 11 , wherein a second predetermined deep learning model is run on the results of the first search for determining the similarity scores for the extracted sentences, wherein the second predetermined deep learning model is a Term Frequency-Inverse Document Frequency (TF-IDF) model.
13 . The CPP of claim 8 , wherein the first descriptive name is selected from the group consisting of: a column name in a database, a header in a comma-separated values (CSV) formatted text file, and a categorical variable in a CSV formatted text file.
14 . The CPP of claim 8 , wherein the first knowledge base is an open and editable knowledge base accessible on the Internet.
15 . A computer system (CS), the CS comprising:
a processor set;
a set of one or more computer-readable storage media;
program instructions, collectively stored in the set of one or more storage media, for causing the processor set to perform the following computer operations:
cause a first search to be performed on a first knowledge base for a first descriptive name;
extract sentences from results of the first search,
wherein the extracted sentences include labels associated with the first descriptive name and descriptions of the labels;
run at least one predetermined deep learning model on the results of the first search for determining similarity scores for the extracted sentences, wherein each of the similarity scores defines a similarity score for an associated one of the extracted sentences and a first glossary;
determine a first of the determined similarity scores, wherein the first determined similarity score has a relatively greater similarity score than the other determined similarity scores; and
use the first determined similarity score to map terms of a second glossary to terms of the first glossary to enhance context provided in search results generated using the glossaries.