US 2022/0101860 A1 · AUTOMATED SPEECH GENERATION BASED ON DEVICE FEED

Abstract

Computer-generated speech based on a device feed includes generating a corpus for robotic use by receiving first information representative of a user's speech in different environments at different times and second information representative of environmental conditions of different locations at the different times. The first and second information of corresponding different environments and different locations for each of the different times is combined with third information received from external data sources. A plurality of annotated combined datasets including the first information, the second information, and the third information is generated for each of the different times in a repository. The plurality of annotated combined datasets is correlated to create training data that is subsequently processed using a predetermined machine learning model. A correlation among spoken tone associated with a contextual situation based on skills of the user is identified in the training data and used to update the corpus.

Claims20

Dependent claims

A computer-implemented method for speech generation, comprising:

generating a corpus for robotic use by:
receiving, by one or more processors, first information representative of user's speech in different locations at different times;
receiving, by the one or more processors, second information representative of environmental conditions of the different locations at the different times;
combining, by the one or more processors, the first information and the second information of corresponding different locations for each of the different times;
in response to receiving third information from external data sources, generating, by the one or more processors, a plurality of annotated combined datasets comprising the first information, the second information, and the third information for each of the different times in a repository;
correlating, by the one or more processors, the plurality of annotated combined datasets to create training data;
and
processing, by the one or more processors, the training data using a predetermined machine learning model;
analyzing, by the one or more processors, the training data to identify a correlation among spoken tone associated with a contextual situation based on skills of a user;
and
updating, by the one or more processors, the corpus with the identified correlation.

8 more elements

A computer system for speech generation, comprising:

one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, wherein the computer system is capable of performing a method comprising:
generating a corpus for robotic use by:
receiving, by one or more processors, first information representative of user's speech in different locations at different times;
receiving, by the one or more processors, second information representative of environmental conditions of the different locations at the different times;
combining, by the one or more processors, the first information and the second information of corresponding different locations for each of the different times;
in response to receiving third information from external data sources, generating, by the one or more processors, a plurality of annotated combined datasets comprising the first information, the second information, and the third information for each of the different times in a repository;
correlating, by the one or more processors, the plurality of annotated combined datasets to create training data;
and
processing, by the one or more processors, the training data using a predetermined machine learning model;
analyzing, by the one or more processors, the training data to identify a correlation among spoken tone associated with a contextual situation based on skills of a user;
and
updating, by the one or more processors, the corpus with the identified correlation.

9 more elements

A computer program product for speech generation, comprising:

one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising:
generating a corpus for robotic use by:
receiving, by one or more processors, first information representative of user's speech in different locations at different times;
receiving, by the one or more processors, second information representative of environmental conditions of the different locations at the different times;
combining, by the one or more processors, the first information and the second information of corresponding different locations for each of the different times;
in response to receiving third information from external data sources, generating, by the one or more processors, a plurality of annotated combined datasets comprising the first information, the second information, and the third information for each of the different times in a repository;
correlating, by the one or more processors, the plurality of annotated combined datasets to create training data;
and
processing, by the one or more processors, the training data using a predetermined machine learning model;
analyzing, by the one or more processors, the training data to identify a correlation among spoken tone associated with a contextual situation based on skills of a user;
and
updating, by the one or more processors, the corpus with the identified correlation.

9 more elements

Dependencies are inferred from the claim text.

Automated speech generation based on device feed

Life of this patent

Automated speech generation based on device feed

Life of this patent

Abstract

Claims20