Methods and systems for automatic queuing in conference calls

Patent metadata

Field	Value
Publication number	US-12052391-B2
Application number	US-202017083241-A
Country	US
Kind code	B2
Filing date	Oct 28, 2020
Priority date	Oct 28, 2020
Publication date	Jul 30, 2024
Grant date	Jul 30, 2024

Abstract

Disclosed are methods, systems, and non-transitory computer-readable medium for automatically queuing participants during a conference call. For instance, the method may include receiving call data associated with a conference call; analyzing the call data to identify the participants on the conference call; and determining whether two or more participants of the plurality of participants are speaking at a same time. The method can also include tracking a first participant that continues speaking and a second participant that stops speaking; and displaying a queuing element on a graphical user interface (GUI) to indicate that the second participant is in a queue to speak once the first participant has finished speaking.

First claim

What is claimed is:

1. A computer-implemented method, the computer-implemented method comprising:
receiving call data associated with a conference call;
analyzing the call data to identify a plurality of participants on the conference call and a call topic;
determining whether two or more participants of the plurality of participants are speaking at a same time;
based upon a determination that the two or more participants are speaking at the same time, tracking the two or more participants to identify a first participant that continues speaking and a second participant that voluntarily yields to the first participant;
determining that the yielding second participant is more relevant to the call topic than the first participant;
muting the first participant based on determining that the yielding second participant is more relevant to the call topic;
displaying a queuing element on a graphical user interface (GUI) associated with the conference call to indicate the two or more participants are speaking at the same time and that the first participant is in a queue to speak after the second participant has finished;
determining when the second participant has stopped speaking by:
extracting an audio portion of the call data;
processing the audio portion to determine text by a speech-to-text function;
processing the text to form text feature vectors; and
determining the second participant shifted from the first call topic to a second call topic based on one or more topic modeling neural network models;

muting audio input of devices of all participants other than the first participant; and
displaying an indicator that it is the first participant's turn to speak.

2. The computer-implemented method of claim 1 , wherein analyzing the call data to identify a plurality of participants on the conference call includes:
extracting a second audio portion of the call data;
processing the second audio portion to form a second feature vector;
processing the second feature vector through the one or more topic modeling neural network models to map utterances to one or more entities; and
mapping the one or more entities to the plurality of participants.

3. The computer-implemented method of claim 1 , wherein analyzing the call data to identify a plurality of participants on the conference call includes:
identifying one or more devices connected to the conference call;
extracting a second audio portion of the call data;
processing the second audio portion to associate utterances to the one or more devices; and
mapping the one or more devices to the plurality of participants.

4. The computer-implemented method of claim 1 , wherein the GUI, as displayed on a device associated with the first participant, includes an opt-out element that when selected by a user input removes the first participant from the queue.

5. The computer-implemented method of claim 1 , wherein determining when the second participant has stopped speaking includes:
determining whether the text includes keywords; and
based upon a determination that the text includes the keywords, determining the second participant has stopped speaking.

6. The computer-implemented method of claim 1 , wherein determining whether the two or more participants are speaking at the same time includes:
determining whether different audio inputs from different devices are input at the same time or determining that there are two or more voices being input at the same time; and
based upon a determination that the different audio inputs are input from the different devices at the same time or that there are two or more voices being input at the same time, determining two or more participants are speaking at the same time.

7. The computer-implemented method of claim 6 , wherein determining whether two or more participants are speaking at the same time further includes, before determining the two or more participants are speaking at the same time:
determining whether an audio input of the different audio inputs is a background noise; and
based on a determination that the audio input of the different audio inputs is not background noise, determining two or more participants are speaking at the same time.

8. The computer-implemented method of claim 6 , wherein determining whether two or more participants are speaking at the same time further includes, before determining two or more participants are speaking at the same time:
determining whether a voice of the two or more voices is a non-verbal utterance; and
based on a determination that the voice is not a non-verbal utterance, determining two or more participants are speaking at the same time.

9. The computer-implemented method of claim 1 , wherein the received call data includes predetermined words; and
determining that the yielding second participant is more relevant to the call topic than the first participant includes comparing the speaking of the first and second participant to the predetermined words.

10. The computer-implemented method of claim 1 , wherein determining the second participant shifted from the first call topic to a second call topic is based on determining a transition point based on the one or more topic modeling neural network models.

11. A system, the system comprising:
a memory storing instructions; and
a processor executing the instructions to perform a process including:
receiving call data associated with a conference call;
analyzing the call data to identify a plurality of persons on the conference call and a call topic;
determining whether two or more persons of the plurality of persons are speaking at a same time;
based upon a determination that the two or more persons are speaking at the same time, tracking the two or more persons to identify a first person that continues speaking and a second person that voluntarily yields to the first person;
determining that the yielding second person is more relevant to the call topic than the first person;
muting the first person based on determining that the yielding second person is more relevant to the call topic;
displaying a queuing element on a graphical user interface (GUI) associated with the conference call to indicate the two or more persons are speaking at the same time and that the first person is in a queue to speak after the second person has finished;
determining when the second person has stopped speaking by:
extracting an audio portion of the call data;
processing the audio portion to determine text by a speech-to-text function:
processing the text to form text feature vectors; and
determining the second person shifted from the first call topic to a second call topic based on one or more topic modeling neural network models;

muting audio input of devices of all participants other than the first person; and
displaying an indicator that it is the first person's turn to speak.

12. The system of claim 11 , wherein analyzing the call data to identify a plurality of persons on the conference call includes:
extracting a second audio portion of the call data;
processing the second audio portion to form a second feature vector;
processing the second feature vector through the one or more topic modeling neural network models to map utterances to one or more entities; and
mapping the one or more entities to the plurality of persons.

13. The system of claim 11 , wherein the process further includes, for analyzing the call data to identify a plurality of persons on the conference call:
identifying one or more devices connected to the conference call;
extracting a second audio portion of the call data;
processing the second audio port

Assignees

• Capital One Services Llc

Inventors

• Pham Vincent
• Goodsitt Jeremy
• Key Kate
• Truong Anh
• Walters Austin

Classifications

• G10L17/22

  Interactive procedures; Man-machine interfaces · CPC title

• G10L17/18

  Artificial neural networks; Connectionist approaches · CPC title

• G10L17/02

  Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction · CPC title

• G06F3/167

  Audio in a user interface, e.g. using voice commands for navigating, audio feedback · CPC title

• G06F3/165

  Management of the audio stream, e.g. setting of volume, audio stream path · CPC title