US 11,893,503 B2Granted

Machine learning based semantic structural hole identification

AssigneeAccenture Global Solutions Limited

InventorsJanardan Misra, Sanjay Podder

Life of this patent

Estimating the term…

Loading patent…

Abstract

In some examples, machine learning based semantic structural hole identification may include mapping each text element of a plurality of text elements of a corpus into an embedding space that includes embeddings that are represented as vectors. A semantic network may be generated based on semantic relatedness between each pair of vectors. A boundary enclosure of the embedding space may be determined, and points to fill the boundary enclosure may be generated. Based on an analysis of voidness for each point within the boundary enclosure, a set of void points and void regions may be identified. Semantic holes may be identified for each void region, and utilized to determine semantic porosity of the corpus. A performance impact may be determined between utilization of the corpus to generate an application by using the text elements without filling the semantic holes and the text elements with the semantic holes filled.

Claims18

Dependent claims

A machine learning based semantic structural hole identification apparatus comprising:

at least one hardware processor;
an embeddings generator, executed by the at least one hardware processor, to map text elements of a corpus into an embedding space that includes a plurality of embeddings that are represented as vectors;
a semantic network generator, executed by the at least one hardware processor, to generate, based on semantic relatedness between each pair of the vectors of the embedding space, a semantic network;
a semantic space hole identifier, executed by the at least one hardware processor, to identify semantic holes in the semantic network;
a data porosity analyzer, executed by the at least one hardware processor, to determine, based on a number of the semantic holes in comparison to the text elements of the corpus, semantic porosity of the corpus;
a void filler, executed by the at least one hardware processor, to fill the semantic holes;
a porosity impact analyzer, executed by the at least one hardware processor, to determine, based on the semantic porosity of the corpus, a performance impact of utilization of the corpus to generate an application by using the text elements of the corpus without filling the semantic holes;
and
an application generator, executed by the at least one hardware processor, to
determine whether the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to an impact threshold;
and
based on a determination that the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to the impact threshold, utilize the text elements of the corpus with the semantic holes filled by the void filler to generate the application.

9 more elements

A computer-implemented method for machine learning based semantic structural hole identification, the method comprising:

generating, by at least one hardware processor, a semantic network from a corpus of text elements;
identifying semantic holes in the semantic network;
based on a number of the semantic holes in comparison to the text elements of the corpus, determining semantic porosity of the corpus;
based on the semantic porosity of the corpus, determining a performance impact of utilization of the corpus to generate an application by using the text elements without filling the semantic holes;
determining whether the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to an impact threshold;
and
based on a determination that the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to the impact threshold, filling the semantic holes and utilizing the text elements of the corpus with the semantic holes filled to generate the application.

4 more elements

A non-transitory computer readable medium having stored thereon machine readable instructions, the machine readable instructions, when executed by at least one hardware processor, cause the at least one hardware processor to:

generate a semantic network for a corpus of text elements;
identify semantic holes in the semantic network;
based on a number of the semantic holes in comparison to the text elements of the corpus, determine semantic porosity of the corpus;
based on the semantic porosity of the corpus, determine a performance impact of utilization of the corpus to generate an application by using the text elements without filling the semantic holes;
determine whether the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to an impact threshold;
and
based on a determination that the performance impact of utilization of the corpus to generate the application without filling the semantic holes is greater than or equal to the impact threshold, fill the semantic holes and utilize the text elements of the corpus with the semantic holes filled to generate the application.

4 more elements

Dependencies are inferred from the claim text.