Natural language processing techniques for machine-learning-guided summarization using hybrid class templates
As described herein, various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing natural language processing operations for generating guided summaries using summarization templates that are mapped to hybrid classes of a hybrid classification space for a hybrid classification machine learning model. In some embodiments, by using summarization templates, a proposed summarization framework is able to vastly reduce the computational complexity of performing summarization on an input document data object, such as an input multi-party communication transcript data object, by defining the set of dynamic data fields that apply to the input document data object based at least in part on an assigned class/category of the input document data object.
1. A computer-implemented method for generating a guided summary for a multi-party communication transcript data object, the computer-implemented method comprising:
generating, using one or more processors and based at least in part on the multi-party communication transcript data object and using a hybrid space classification machine learning model, a hybrid class for the multi-party communication transcript data object that comprises a primary class for the multi-party communication transcript data object and a secondary class for the multi-party communication transcript data object, wherein: (i) the primary class is selected from a primary classification space, and (ii) the secondary class is selected from a secondary classification space that is distinct from the primary classification space;
generating, using the one or more processors and based at least in part on the hybrid class, a summarization template for the multi-party communication transcript data object, wherein the summarization template defines: (i) a plurality of template static text segments associated with the hybrid class, and (ii) a plurality of template dynamic data fields;
generating, using the one or more processors and based at least in part on monitored user activity data for a monitored end user associated with the multi-party communication transcript data object during a communication time period, one or more predicted data field values for the plurality of template dynamic data fields;
generating, using the one or more processors and based at least in part on the summarization template and the one or more predicted data field values, the guided summary; and
performing, using the one or more processors, one or more prediction-based actions based at least in part on the guided summary.
2. The computer-implemented method of claim 1 , wherein:
the primary classification space comprises a plurality of defined primary classes, and
each defined primary class of the plurality of defined primary classes is associated with a respective distinct secondary classification space that comprises a plurality of sub-classes of the defined primary class.
3. The computer-implemented method of claim 1 , wherein the monitored user activity data is generated using a user activity monitoring software that executes on an application server computing entity that the monitored end user interacts with via a networked connection with an end user computing entity and is extracted via a monitored user activity data reporting application programming interface (API) that is executed by the application server computing entity.
4. The computer-implemented method of claim 3 , wherein each template dynamic data field of the plurality of template dynamic data fields is mapped to a schema segment of an API schema of the monitored user activity data reporting API using API response mapping data.
5. The computer-implemented method of claim 1 , wherein:
the hybrid class is selected from a hybrid classification space that comprises a group of defined hybrid classes, and
each defined hybrid class of the group of defined hybrid classes is associated with a respective distinct summarization template of a group of distinct summarization templates.
6. The computer-implemented method of claim 1 , wherein generating a respective distinct summarization template for a particular defined hybrid class comprises:
identifying a plurality of qualified historical multi-party communication documentation data objects for the particular defined hybrid class, wherein each qualified historical multi-party communication documentation data object of the plurality of qualified historical multi-party communication documentation data objects: (i) is predicted to be associated with the particular defined hybrid class using the hybrid space classification machine learning model, and (ii) satisfies one or more documentation data object criteria;
generating, based at least in part on the plurality of qualified historical multi-party communication documentation data objects, a plurality of repeating segments and a plurality of high variability segments;
generating the plurality of template static text segments in the respective distinct summarization template based at least in part on the plurality of repeating segments; and
generating the plurality of template dynamic data fields in the respective distinct summarization template based at least in part on the plurality of high variability segments.
7. The computer-implemented method of claim 6 , wherein the one or more documentation data object criteria comprise a length criterion.
8. The computer-implemented method of claim 6 , wherein the one or more documentation data object criteria comprise a keyword presence criterion.
9. A system comprising one or more processors and at least one memory storing processor-executable instructions that, when executed by any one or more of the one or more processors, cause the one or more processors to perform operations comprising:
generating, based at least in part on a multi-party communication transcript data object and using a hybrid space classification machine learning model, a hybrid class for the multi-party communication transcript data object that comprises a primary class for the multi-party communication transcript data object and a secondary class for the multi-party communication transcript data object, wherein: (i) the primary class is selected from a primary classification space, and (ii) the secondary class is selected from a secondary classification space that is distinct from the primary classification space;
generating, based at least in part on the hybrid class, a summarization template for the multi-party communication transcript data object, wherein the summarization template defines: (i) a plurality of template static text segments associated with the hybrid class, and (ii) a plurality of template dynamic data fields;
generating, based at least in part on monitored user activity data for a monitored end user associated with the multi-party communication transcript data object during a communication time period, one or more predicted data field values;
generating, based at least in part on the summarization template and the one or more predicted data field values, a guided summary; and
performing one or more prediction-based actions based at least in part on the guided summary.
10. The system of claim 9 , wherein:
the primary classification space comprises a plurality of defined primary classes, and
each defined primary class of the plurality of defined primary classes is associated with a respective distinct secondary classification space that comprises a plurality of sub-classes of the defined primary class.
11. The system of claim 9 , wherein the monitored user activity data is generated using a user activity monitoring software that executes on an application server computing entity that the monitored end user interacts with via a networked connection with an end user computing entity and is extracted via a monitored user activity data reporting application programming interface (API) that is executed by the application server computing entity.
12. The system of claim 11 , wherein each template dynamic data field of the plurality of template dynamic data fields is mapped to a schema segment of an API schema of the monitored user activity data reporting API using API response mapping data.
13. The system of claim 9 , wherein:
the hybrid class is selected from a hybrid classification space that comprises a group of defined hybrid classes, and
each defined hybrid class of the group of defined hybrid classes is associated with a respective distinct summarization template of a group of distinct summarization templates.
14. The system of claim 9 , wherein generating a respective distinct summarization template for a particular defined hybrid class comprises:
identifying a plurality of qualified historical multi-party communication documentation data objects for the particular defined hybrid class, wherein each qualified historical multi-party communication documentation data object of the plurality of qualified historical multi-party communication documentation data objects: (i) is predicted to be associated with the particular defined hybrid class using the hybrid space classification machine learning model, and (ii) satisfies one or more documentation data object criteria;
generating, based at least in part on the plurality of qualified historical multi-party communication documentation data objects, a plurality of repeating segments and a plurality of high variability segments;
generating the plurality of template static text segments in the respective distinct summarization template based at least in part on the plurality of repeating segments; and
generating the plurality of template dynamic data fields in the respective distinct summarization template based at least in part on the plurality of high variability segments.
15. The system of claim 14 , wherein the one or more documentation data object criteria comprise a length criterion.
16. The system of claim 14 , wherein the one or more documentation data object criteria comprise a keyword presence criterion.
17. One or more non-transitory computer-readable storage media storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
generating, based at least in part on a multi-party communication transcript data object and using a hybrid space classification machine learning model, a hybrid class for the multi-party communication transcript data object that comprises a primary class for the multi-party communication transcript data object and a secondary class for the multi-party communication transcript data object, wherein: (i) the primary class is selected from a primary classification space, and (ii) the secondary class is selected from a secondary classification space that is distinct from the primary classification space;
generating, based at least in part on the hybrid class, a summarization template for the multi-party communication transcript data object, wherein the summarization template defines: (i) a plurality of template static text segments associated with the hybrid class, and (ii) a plurality of template dynamic data fields;
generating, based at least in part on monitored user activity data for a monitored end user associated with the multi-party communication transcript data object during a communication time period, one or more predicted data field values;
generating, based at least in part on the summarization template and one or more predicted data field values, a guided summary; and
performing one or more prediction-based actions based at least in part on the guided summary.
18. The one or more non-transitory computer-readable storage media of claim 17 , wherein:
the primary classification space comprises a plurality of defined primary classes, and
each defined primary class of the plurality of defined primary classes is associated with a respective distinct secondary classification space that comprises a plurality of sub-classes of the defined primary class.
19. The one or more non-transitory computer-readable storage media of claim 17 , wherein the monitored user activity data is generated using a user activity monitoring software that executes on an application server computing entity that the monitored end user interacts with via a networked connection with an end user computing entity and is extracted via a monitored user activity data reporting application programming interface (API) that is executed by the application server computing entity.
20. The one or more non-transitory computer-readable storage media of claim 19 , wherein each template dynamic data field of the plurality of template dynamic data fields is mapped to a schema segment of an API schema of the monitored user activity data reporting API using API response mapping data.