Communication tone training

Patent metadata

Field	Value
Publication number	US-10832587-B2
Application number	US-201715459561-A
Country	US
Kind code	B2
Filing date	Mar 15, 2017
Priority date	Mar 15, 2017
Publication date	Nov 10, 2020
Grant date	Nov 10, 2020

Abstract

An approach is provided that may obtain communication information regarding a communication between a first entity and a second entity while the communication may be ongoing. The communication may include an utterance. A tone associated with the utterance may be identified and may result in an identified tone. An outcome of the communication may be predicted using a machine learning based prediction module and the identified tone.

First claim

The invention claimed is:

1. A computer program product stored in a computer readable storage medium, comprising computer program code that, when executed by an information handling system, causes the information handling system to provide communication tone training by performing actions comprising:
obtaining, by a communication tone training monitor using machine-learning that includes at least one processor, communication information regarding a communication between a first entity and a second entity while the communication is ongoing, wherein the communication includes an utterance;
parsing, by the communication tone training module monitor, the communication information to extract a tone associated with the utterance, resulting in an identified tone in a dialog flow, wherein the dialog flow includes a plurality of identified tones and communication turns for a period of time that describe a tonal evolution of the communication between the first entity and the second entity resulting in a tonal pattern of the communication that may be analyzed in context;
predicting, by the communication tone training monitor, an outcome of the communication by accessing a tonal patterns repository and determining whether the tonal pattern of the communication complies with a best-practice tonal pattern from the tonal patterns repository, wherein the tonal patterns repository is operatively coupled to a plurality of entities, and wherein the tonal patterns repository includes a plurality of best-practice tonal patterns based on tones and communication turns for a period of time, resulting in a prediction score;
determining whether the prediction score is below a threshold;
when the prediction score is below the threshold, flagging the communication;
when the prediction score is below the threshold, determining an intervention plan intended
to increase the prediction score and outputting the intervention plan;
using, by the communication tone training monitor, the prediction score to determine guidance regarding tonal choice to the second entity for a next turn of the communication, wherein the tonal choice is selected from one of the detected tones from one of the best-practice tonal patterns in the tonal patterns repository; and
outputting the guidance to a display.

2. The computer program product of claim 1 , wherein the prediction score is below a threshold, and further comprising computer program code that, when executed by the information handling system, causes the information handling system to perform actions further comprising:
providing a sample sentence to the second entity to use for a next turn of communication;
outputting an alert that assistance is needed;
annotating the communication information to associate the identified tone with the utterance and to indicate that the dialog flow fails to comply with any of the best-practice tonal patterns in the tonal patterns repository, resulting in annotated communication information; and
providing the annotated communication information for updating the tonal patterns repository.

3. The computer program product of claim 1 , and further comprising computer program code that, when executed by the information handling system, causes the information handling system to perform actions further comprising:
identifying additional tones associated with additional utterances, resulting in additional identified tones in the dialog flow;
annotating the communication information to associate the additional identified tones with the additional utterances to indicate a tonal evolution of the communication;
determining whether the tonal evolution of the communication complies with any of the best-practice tonal patterns in the tonal patterns repository; and
if it is determined that the tonal evolution of the communication fails to comply with any of the best-practice tonal patterns in the tonal patterns repository, flagging the communication.

4. The computer program product of claim 1 , and further comprising computer program code that, when executed by the information handling system, causes the information handling system to perform actions further comprising:
determining whether the identified tone is inappropriate for a context of the communication; and
if it is determined that the identified tone is inappropriate, flagging the communication.

5. The computer program product of claim 1 , and further comprising computer program code that, when executed by the information handling system, causes the information handling system to perform actions further comprising:
suggesting training to improve a future communication.

6. An information handling system comprising:
one or more processors;
one or more data stores accessible by at least one processor;
a memory coupled to at least one of the processors; and
a set of computer program instructions stored in the memory and executed by at least one of the processors to perform the actions of:
obtaining, by a communication tone training monitor using machine learning, communication information regarding a communication between a first entity and a second entity while the communication is ongoing, wherein the communication includes an utterance;
parsing, by the communication tone training monitor, the communication information to extract a tone associated with the utterance, resulting in an identified tone in a dialog flow, wherein the dialog flow includes a plurality of identified tones and communication turns for a period of time that describe a tonal evolution of the communication between the first entity and the second entity resulting in a tonal pattern of the communication that may be analyzed in context;
predicting, by the communication tone training monitor, an outcome of the communication by accessing a tonal patterns repository and determining whether the tonal pattern of the communication complies with a best-practice tonal pattern from the tonal patterns repository, wherein the tonal patterns repository is operatively coupled to a plurality of entities, and wherein the tonal patterns repository includes a plurality of best-practice tonal patterns based on tones and communication turns for a period of time, resulting in a prediction score;
determining whether the prediction score is below a threshold;
when the prediction score is below the threshold, flagging the communication;
when the prediction score is below the threshold, determining an intervention plan intended to increase the prediction score and outputting the intervention plan;
using, by the communication tone training monitor, the prediction score to determine guidance regarding tonal choice to the second entity for the next turn of communication, wherein the tonal choice is selected from one of the detected tones from one of the best-practice tonal patterns in the tonal patterns repository; and
outputting the guidance to a display.

7. The information handling system of claim 6 , wherein the prediction score is below a threshold, and wherein the set of computer program instructions further comprises instructions executed by at least one of the processors to perform the actions of:
providing a sample sentence to the second entity to use for a next turn of communication;
outputting an alert that assistance is needed;
annotating the communication information to associate the identified tone with the utterance and to indicate that the dialog flow fails to comply with any of the best-practice tonal patterns in the tonal patterns repository, resulting in annotated communication information; and
providing the annotated communication information for updating the tonal patterns repository.

8. The information handling system of claim 6 , wherein the set of computer program instruct

Assignees

• IBM

Inventors

• Akkiraju Rama K
• Mahmud Jalal U
• Sinha Vibha S
• Zhang Mengdi

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• G09B19/00 Primary

  Teaching not covered by other main groups of this subclass (teaching or practice apparatus for gun-aiming or gun-laying F41G3/26) · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G09B5/06

  with both visual and audible presentation of the material to be studied · CPC title

• G06F40/253

  Grammatical analysis; Style critique · CPC title