Learning framework for processing communication session transcripts

Patent metadata

Field	Value
Publication number	US-2025148208-A1
Application number	US-202519012092-A
Country	US
Kind code	A1
Filing date	Jan 7, 2025
Priority date	Dec 17, 2021
Publication date	May 8, 2025
Grant date	—

Abstract

Learning frameworks for processing text transcripts may include receiving, by an application, a query comprising a topic. A topic model may determine a plurality of subtopics based on the topic. The application may receive, from a database based on the topic and plurality of subtopics, a plurality of text transcripts. A sentiment model may compute, for each text transcript, a respective sentiment score based on a text of the respective text transcript. The application may determine, for each text transcript, a duration of a communication session associated with the respective text transcript. The application may compute, for each text transcript, a total score based on the sentiment score and the duration of the respective text transcript. The application may return, as responsive to the query, a subset of the plurality of text transcripts having a total score that exceeds a threshold.

First claim

What is claimed is:

1 . A computing apparatus comprising:
a processing circuit; and a memory having executable instructions stored thereon, which when executed by the processing circuit, cause the processing circuit to:
receive a plurality of transcripts, each associated with a corresponding communication session;
process the plurality of transcripts to identify a topic of each of the corresponding communication sessions;
output the identified topics of the corresponding communication sessions for display on a graphical user interface (GUI) of a search application operating on the computing apparatus;
receive a selection of at least one of the identified topics;
execute the search application to generate and submit a query including the selection of the at least one identified topics;
output one or more of the plurality of transcripts for the corresponding communication sessions that match the selection;
extract a plurality of key phrases from each of the one or more transcripts;
determine, based on content of the plurality of key phrases, a system or network outage; and
send instructions to a computing device in communication with the system for restoring the system or network outage.

2 . The computing apparatus of claim 1 , wherein each of the corresponding communication sessions is a phone call, online text-based chat session, or text-based communication session.

3 . The computing apparatus of claim 1 , wherein at least one of the plurality of transcripts includes notes taken by an agent for the corresponding communication session, wherein the processor is further caused to display the notes on the GUI in a notes column on a row for the corresponding identified topic being output.

4 . The computing apparatus of claim 1 , wherein the query further includes one or more subtopics related to the at least one identified topics.

5 . The computing apparatus of claim 4 , wherein the one or more of the plurality of transcripts that are output further match the one or more subtopics.

6 . The computing apparatus of claim 1 , wherein the plurality of key phrases is extracted using a key phrase model that has been trained based on a plurality of training transcripts using unsupervised training.

7 . The computing apparatus of claim 1 , wherein the processing circuit is further caused to:
assign a score to each of the plurality of transcripts, the score for a corresponding transcript being assigned based on a sentiment model analysis of the corresponding transcript and a duration of the communication session of the corresponding transcript; wherein the sentiment model includes a natural language processing model configured to detect a sentiment of a party to the communication session; and wherein the score reflects whether a negative sentiment is detected by the sentiment model.

8 . A method comprising:
processing a plurality of transcripts, each transcript being associated with a corresponding communication session, to identify a topic of each of the corresponding communication sessions; causing the identified topics of the corresponding communication sessions to be displayed; processing a selection of at least one of the identified topics; identifying a search application to generate and submit a query including the selection of the at least one identified topics; outputting one or more of the plurality of transcripts for the corresponding communication sessions that match the selection; processing a plurality of key phrases from each of the one or more transcripts; determining based on content of the plurality of key phrases a system or network outage; and sending instructions to a computing device in communication with the system for restoring the system or network outage.

9 . The method of claim 8 , wherein each of the corresponding communication sessions is a phone call, online text-based chat session, or text-based communication session.

10 . The method of claim 8 , wherein at least one of the plurality of transcripts includes notes taken by an agent for the corresponding communication session, wherein the processor is further caused to display the notes on the GUI in a notes column on a row for the corresponding identified topic being output.

11 . The method of claim 8 , wherein the query further includes one or more subtopics related to the at least one identified topics.

12 . The method of claim 11 , wherein the one or more of the plurality of transcripts that are output further match the one or more subtopics.

13 . The method of claim 8 , wherein the plurality of key phrases is extracted using a key phrase model that has been trained based on a plurality of training transcripts using unsupervised training.

14 . The method of claim 8 , further comprising:
assigning a score to each of the plurality of transcripts, the score for a corresponding transcript being assigned based on a sentiment model analysis of the corresponding transcript and a duration of the communication session of the corresponding transcript; wherein the sentiment model includes a natural language processing model configured to detect a sentiment of a party to the communication session; and wherein the score reflects whether a negative sentiment is detected by the sentiment model.

15 . A non-transitory computer-readable storage medium having executable instructions stored thereon, which when executed by a processor, cause the processor to:
process a plurality of transcripts, each of the plurality of transcripts being associated with a corresponding communication session, to identify a topic of each of the corresponding communication sessions; cause the identified topics of the corresponding communication sessions to be displayed; identify a selection of at least one of the identified topics; generate and submit a query including the selection of the at least one identified topics; output one or more of the plurality of transcripts for the corresponding communication sessions that match the selection; extract a plurality of key phrases from each of the one or more transcripts; determine based on content of the plurality of key phrases a system or network outage; and send instructions to a computing device in communication with the system for restoring the system or network outage.

16 . The computer-readable storage medium of claim 15 , wherein each of the corresponding communication sessions is a phone call, online text-based chat session, or text-based communication session.

17 . The computer-readable storage medium of claim 15 , wherein at least one of the plurality of transcripts includes notes taken by an agent for the corresponding communication session, wherein the processor is further caused to display the notes on the GUI in a notes column on a row for the corresponding identified topic being output.

18 . The computer-readable storage medium of claim 15 , wherein the query further includes one or more subtopics related to the at least one identified topics; and
wherein the one or more of the plurality of transcripts that are output further match the one or more subtopics.

19 . The computer-readable storage medium of claim 15 , wherein the plurality of key phrases is extracted using a key phrase model that has been trained based on a plurality of training transcripts using unsupervised training.

20 . The computer-readable storage medium of claim 15 , wherein the processing circuit is further caused to:
assign a score to each of the plurality of transcripts, th

Assignees

• Capital One Services Llc

Inventors

• Thoniparambil Akshay Aravindakshan
• Agarwal Manish
• Kedia Gaurav
• Bansiwala Aashish
• Ramamohan Deepak
• Choudhary Sourav

Classifications

• G06F40/35

  Discourse or dialogue representation · CPC title

• G06F16/3344

  using natural language analysis · CPC title

• G06F16/355

  Creation or modification of classes or clusters · CPC title

• G06F16/338

  Presentation of query results · CPC title

• G06F40/279

  Recognition of textual entities · CPC title