Voice based activation detection

1 . A computer-implemented method (CIM), comprising:

in response to receiving an audio signal having initial utterances from a user, determining whether the initial utterances include predetermined wake-up words by:

using a trained artificial intelligence (AI) based model to evaluate a tone of the initial utterances;

in response to determining that the initial utterances include the predetermined wake-up words, causing a voice assistant to initiate an interaction with the user;

in response to receiving additional utterances from the user that do not have an agitated tone, causing the additional utterances to be processed;

in response to determining that causing the additional utterances to be processed results in action being taken, labeling the initial utterances that include the wake-up words as genuine dialog initiation entries;

in response receiving additional utterances from the user that do have an agitated tone, labeling the initial utterances that include the wake-up words as unintentional dialog initiation entries;

causing the additional utterances to be ignored; and

using the labeled unintentional dialog initiation entries to retrain the AI based model.

2 . The CIM of claim 1 , further comprising:

using the labeled genuine dialog initiation entries to retrain the AI based model.

3 . The CIM of claim 1 , further comprising:

in response to determining that the initial utterances in the audio signal do not stop for a predetermined period following the initial utterances that include the wake-up words, labeling the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

using the labeled unintentional dialog initiation entries to retrain the AI based model.

4 . The CIM of claim 1 , further comprising:

in response to determining that additional utterances are not received from the user following a predetermined period, labeling the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

using the labeled unintentional dialog initiation entries to retrain the AI based model.

5 . The CIM of claim 1 , further comprising:

in response to determining that causing the additional utterances to be processed results in no action being taken, labeling the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

using the labeled unintentional dialog initiation entries to retrain the AI based model.

6 . The CIM of claim 1 , wherein the trained AI based model is a trained machine learning model.

7 . The CIM of claim 1 , further comprising:

receiving a mixed audio signal having initial utterances from multiple users;

extracting individual audio signals from the mixed audio signal, each of the extracted individual audio signals having ones of the initial utterances that correspond to a respective one of the multiple users; and

for each of the individual audio signals, determining whether the ones of the initial utterances in a given individual audio signal include the predetermined wake-up words.

8 . A computer program product (CPP), comprising:

a set of one or more computer-readable storage media; and

program instructions, collectively stored in the set of one or more storage media, for causing a processor set to perform the following computer operations:

in response to receiving an audio signal having initial utterances from a user, determining whether the initial utterances include predetermined wake-up words by:

using a trained artificial intelligence (AI) based model to evaluate a tone of the initial utterances;

in response to determining that the initial utterances include the predetermined wake-up words, cause a voice assistant to initiate an interaction with the user;

in response to receiving additional utterances from the user that do not have an agitated tone, cause the additional utterances to be processed;

in response to determining that causing the additional utterances to be processed results in action being taken, label the initial utterances that include the wake-up words as genuine dialog initiation entries;

in response to receiving additional utterances from the user that have an agitated tone, label the initial utterances that include the wake-up words as unintentional dialog initiation entries;

cause the additional utterances to be ignored; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.

9 . The CPP of claim 8 , wherein the program instructions are for causing the processor set to further perform the following computer operations:

use the labeled genuine dialog initiation entries to retrain the AI based model.

10 . The CPP of claim 8 , wherein the program instructions are for causing the processor set to further perform the following computer operations:

in response to determining that the initial utterances in the audio signal do not stop for a predetermined period following the initial utterances that include the wake-up words, label the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.

11 . The CPP of claim 8 , wherein the program instructions are for causing the processor set to further perform the following computer operations:

in response to determining that additional utterances are not received from the user following a predetermined period, label the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.

12 . The CPP of claim 8 , wherein the program instructions are for causing the processor set to further perform the following computer operations:

in response to determining that causing the additional utterances to be processed results in no action being taken, label the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.

13 . The CPP of claim 8 , wherein the trained AI based model is a trained machine learning model.

14 . The CPP of claim 8 , wherein the program instructions are for causing the processor set to further perform the following computer operations:

receive a mixed audio signal having initial utterances from multiple users;

extract individual audio signals from the mixed audio signal, each of the extracted individual audio signals having ones of the initial utterances that correspond to a respective one of the multiple users; and

for each of the individual audio signals, determining whether the ones of the initial utterances in a given individual audio signal include the predetermined wake-up words.

15 . A computer system (CS), comprising:

a processor set;

a set of one or more computer-readable storage media;

program instructions, collectively stored in the set of one or more storage media, for causing the processor set to perform the following computer operations:

in response to receiving an audio signal having initial utterances from a user, determining whether the initial utterances include predetermined wake-up words by:

using a trained artificial intelligence (AI) based model to evaluate a tone of the initial utterances;

in response to determining that the initial utterances include the predetermined wake-up words, cause a voice assistant to initiate an interaction with the user;

in response to receiving additional utterances from the user that do not have an agitated tone, cause the additional utterances to be processed;

in response to determining that causing the additional utterances to be processed results in action being taken, label the initial utterances that include the wake-up words as genuine dialog initiation entries;

in response to receiving additional utterances from the user that have an agitated tone, label the initial utterances that include the wake-up words as unintentional dialog initiation entries;

cause the additional utterances to be ignored; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.

16 . The CS of claim 15 ,

wherein the program instructions are for causing the processor set to further perform the following computer operations:

in response to determining that: the initial utterances in the audio signal do not stop for a predetermined period following the initial utterances that include the wake-up words and additional utterances are not received from the user following the predetermined period, label the initial utterances that include the wake-up words as unintentional dialog initiation entries; and

use the labeled unintentional dialog initiation entries to retrain the AI based model.