Automatically determining language for speech recognition of spoken utterance received via an automated assistant interface
Determining a language for speech recognition of a spoken utterance received via an automated assistant interface for interacting with an automated assistant. Implementations can enable multilingual interaction with the automated assistant, without necessitating a user explicitly designate a language to be utilized for each interaction. Implementations determine a user profile that corresponds to audio data that captures a spoken utterance, and utilize language(s), and optionally corresponding probabilities, assigned to the user profile in determining a language for speech recognition of the spoken utterance. Some implementations select only a subset of languages, assigned to the user profile, to utilize in speech recognition of a given spoken utterance of the user. Some implementations perform speech recognition in each of multiple languages assigned to the user profile, and utilize criteria to select only one of the speech recognitions as appropriate for generating and providing content that is responsive to the spoken utterance.
1 . A method implemented by one or more processors, the method comprising:
processing audio data using an acoustic model corresponding to a given language to monitor for an occurrence of an invocation phrase configured to invoke an automated assistant, wherein the audio data is based on detection of spoken input of a user at a client device that includes an automated assistant interface for interacting with the automated assistant;
detecting, based on processing the audio data using the acoustic model, the occurrence of the invocation phrase in a portion of the audio data;
determining, based on processing of the audio data using the acoustic model, that the portion of the audio data that includes the invocation phrase corresponds to a user profile that is accessible to the automated assistant;
identifying a set of languages, including the given language, assigned to the user profile;
selecting a speech recognition model for the given language;
using the selected speech recognition model to process a subsequent portion of the audio data that follows the portion of the audio data;
causing the automated assistant to provide responsive content that is determined based on the processing of the subsequent portion using the selected speech recognition model;
subsequent to causing the automated assistant to provide the responsive content:
processing additional audio data based on detection of additional spoken input of the user;
using the selected speech recognition model to process the additional audio data to generate a first candidate text representation of the additional audio data in the given language;
identifying an additional language, in the set of languages assigned to the user profile;
selecting an additional speech recognition model, corresponding to the additional language;
using the selected additional speech recognition model to process the additional audio data to generate a second candidate text representation of the additional audio data in the additional language;
selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language; and
causing the automated assistant to provide additional responsive content that is determined based on processing the second candidate text representation.
2 . The method of claim 1 , wherein selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language comprises:
identifying one or more contextual parameters associated with the additional audio data; and
selecting the second candidate text representation in the additional language based on the one or more contextual parameters being more strongly associated, in the user profile, with the additional language than with the given language.
3 . The method of claim 2 , wherein the one or more contextual parameters comprise an identifier of the client device.
4 . The method of claim 2 , wherein the one or more contextual parameters comprise one or multiple of: a time of day, a day of the week, and a location of the client device.
5 . The method of claim 1 , wherein the automated assistant is configured to access multiple different user profiles that are: available at the client device, and associated with multiple different users of the client device.
6 . The method of claim 5 , wherein the multiple different user profiles each identify a set of corresponding languages and a corresponding language probability for each of the corresponding languages, the corresponding language probabilities each based on previous interactions between a corresponding one of the multiple different users and the automated assistant.
7 . A non-transitory computer readable storage medium configured to store instructions that, when executed by one or more processors, cause the one or more processors to perform operations that include:
processing audio data using an acoustic model corresponding to a given language to monitor for an occurrence of an invocation phrase configured to invoke an automated assistant, wherein the audio data is based on detection of spoken input of a user at a client device that includes an automated assistant interface for interacting with the automated assistant;
detecting, based on processing the audio data using the acoustic model, the occurrence of the invocation phrase in a portion of the audio data;
determining, based on processing of the audio data using the acoustic model, that the portion of the audio data that includes the invocation phrase corresponds to a user profile that is accessible to the automated assistant;
identifying a set of languages, including the given language, assigned to the user profile;
selecting a speech recognition model for the given language;
using the selected speech recognition model to process a subsequent portion of the audio data that follows the portion of the audio data;
causing the automated assistant to provide responsive content that is determined based on the processing of the subsequent portion using the selected speech recognition model;
subsequent to causing the automated assistant to provide the responsive content:
processing additional audio data based on detection of additional spoken input of the user;
using the selected speech recognition model to process the additional audio data to generate a first candidate text representation of the additional audio data in the given language;
identifying an additional language, in the set of languages assigned to the user profile;
selecting an additional speech recognition model, corresponding to the additional language;
using the selected additional speech recognition model to process the additional audio data to generate a second candidate text representation of the additional audio data in the additional language;
selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language; and
causing the automated assistant to provide additional responsive content that is determined based on processing the second candidate text representation.
8 . The non-transitory computer readable storage medium of claim 7 , wherein selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language comprises:
identifying one or more contextual parameters associated with the additional audio data; and
selecting the second candidate text representation in the additional language based on the one or more contextual parameters being more strongly associated, in the user profile, with the additional language than with the given language.
9 . The non-transitory computer readable storage medium of claim 8 , wherein the one or more contextual parameters comprise an identifier of the client device.
10 . The non-transitory computer readable storage medium of claim 8 , wherein the one or more contextual parameters comprise one or multiple of: a time of day, a day of the week, and a location of the client device.
11 . The non-transitory computer readable storage medium of claim 7 , wherein the automated assistant is configured to access multiple different user profiles that are: available at the client device, and associated with multiple different users of the client device.
12 . The non-transitory computer readable storage medium of claim 11 , wherein the multiple different user profiles each identify a set of corresponding languages and a corresponding language probability for each of the corresponding languages, the corresponding language probabilities each based on previous interactions between a corresponding one of the multiple different users and the automated assistant.
13 . A system comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations that include:
processing audio data using an acoustic model corresponding to a given language to monitor for an occurrence of an invocation phrase configured to invoke an automated assistant, wherein the audio data is based on detection of spoken input of a user at a client device that includes an automated assistant interface for interacting with the automated assistant;
detecting, based on processing the audio data using the acoustic model, the occurrence of the invocation phrase in a portion of the audio data;
determining, based on processing of the audio data using the acoustic model, that the portion of the audio data that includes the invocation phrase corresponds to a user profile that is accessible to the automated assistant;
identifying a set of languages, including the given language, assigned to the user profile;
selecting a speech recognition model for the given language;
using the selected speech recognition model to process a subsequent portion of the audio data that follows the portion of the audio data;
causing the automated assistant to provide responsive content that is determined based on the processing of the subsequent portion using the selected speech recognition model;
subsequent to causing the automated assistant to provide the responsive content:
processing additional audio data based on detection of additional spoken input of the user;
using the selected speech recognition model to process the additional audio data to generate a first candidate text representation of the additional audio data in the given language;
identifying an additional language, in the set of languages assigned to the user profile;
selecting an additional speech recognition model, corresponding to the additional language;
using the selected additional speech recognition model to process the additional audio data to generate a second candidate text representation of the additional audio data in the additional language;
selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language; and
causing the automated assistant to provide additional responsive content that is determined based on processing the second candidate text representation.
14 . The system of claim 13 , wherein selecting the second candidate text representation in the additional language, in lieu of the first candidate text representation in the given language comprises:
identifying one or more contextual parameters associated with the additional audio data; and
selecting the second candidate text representation in the additional language based on the one or more contextual parameters being more strongly associated, in the user profile, with the additional language than with the given language.
15 . The system of claim 13 , wherein the one or more contextual parameters comprise an identifier of the client device.
16 . The system of claim 13 , wherein the one or more contextual parameters comprise one or multiple of: a time of day, a day of the week, and a location of the client device.
17 . The system of claim 13 , wherein the automated assistant is configured to access multiple different user profiles that are: available at the client device, and associated with multiple different users of the client device.
18 . The system of claim 17 , wherein the multiple different user profiles each identify a set of corresponding languages and a corresponding language probability for each of the corresponding languages, the corresponding language probabilities each based on previous interactions between a corresponding one of the multiple different users and the automated assistant.