Automated assistants with conference capabilities
Techniques are described related to enabling automated assistants to enter into a “conference mode” in which they can “participate” in meetings between multiple human participants and perform various functions described herein. In various implementations, an automated assistant implemented at least in part on conference computing device(s) may be set to a conference mode in which the automated assistant performs speech-to-text processing on multiple distinct spoken utterances, provided by multiple meeting participants, without requiring explicit invocation prior to each utterance. The automated assistant may perform semantic processing on first text generated from the speech-to-text processing of one or more of the spoken utterances, and generate, based on the semantic processing, data that is pertinent to the first text. The data may be output to the participants at conference computing device(s). The automated assistant may later determine that the meeting has concluded, and may be set to a non-conference mode.
1. A method implemented by one or more processors, comprising:
setting an automated assistant implemented at least in part on one or more conference computing devices to a conference mode in which the automated assistant performs speech-to-text processing on multiple distinct spoken utterances exchanged during a conversation between multiple participants, without requiring explicit invocation of the automated assistant prior to each of the multiple distinct spoken utterances;
automatically performing, by the automated assistant, semantic processing on first text generated from the speech-to-text processing of a first spoken utterance of the multiple distinct spoken utterances, wherein the semantic processing is performed without explicit participant invocation;
generating, by the automated assistant, based on the semantic processing, a first query;
obtaining first information that is responsive to the first query;
monitoring, by the automated assistant, the conversation for a pause of at least a first predetermined time interval;
in response to detecting, based on the monitoring, the pause of at least the first predetermined time interval in the conversation, providing audible output that conveys at least part of the first information that is responsive to the first query to the multiple participants at one or more of the conference computing devices while the automated assistant is in conference mode;
in response to determining that a second predetermined time interval since the first spoken utterance that was used to generate the first query has passed without the pause being detected, determining that a relevancy score of the first information fails to satisfy a minimum relevancy threshold; and
providing output that conveys second information that is responsive to a second query generated based on a second spoken utterance that occurred subsequent to the first spoken utterance, wherein the second predetermined time interval is greater than the first predetermined time interval.
2. The method of claim 1 , further comprising, in response to determining that the second predetermined time interval has passed without the pause being detected, outputting the first information after a conclusion of the conversation.
3. A system comprising one or more processors and memory storing instructions that, in response to execution of the instructions by the one or more processors, cause the one or more processors to:
set an automated assistant implemented at least in part on one or more conference computing devices to a conference mode in which the automated assistant performs speech-to-text processing on multiple distinct spoken utterances exchanged during a conversation between multiple participants, without requiring explicit invocation of the automated assistant prior to each of the multiple distinct spoken utterances;
automatically perform, by the automated assistant, semantic processing on first text generated from the speech-to-text processing of a first spoken utterance of the multiple distinct spoken utterances, wherein the semantic processing is performed without explicit participant invocation;
generate, by the automated assistant, based on the semantic processing, a first query;
obtain first information that is responsive to the first query;
monitor, by the automated assistant, the conversation for a pause of at least a first predetermined time interval;
in response to detection of the pause of at least the first predetermined time interval in the conversation, provide audible output that conveys at least part of the first information that is responsive to the first query to the multiple participants at one or more of the conference computing devices while the automated assistant is in conference mode;
in response to a determination that a second predetermined time interval since the first spoken utterance that was used to generate the first query has passed without the pause being detected, determine that a relevancy score of the first information fails to satisfy a minimum relevancy threshold; and
provide output that conveys second information that is responsive to a second query generated based on a second spoken utterance that occurred subsequent to the first spoken utterance, wherein the second predetermined time interval is greater than the first predetermined time interval.
4. The system of claim 3 , further comprising instructions to, in response to the determination that the second predetermined time interval has passed without the pause being detected, output the first information after a conclusion of the conversation.