Back to results
US 10,540,971 B2Granted
System and methods for in-meeting group assistance using a virtual assistant
AssigneeBlue Jeans Network, Inc.
InventorNavneet Kumar
Life of this patent
Estimating the term…
Abstract
During a video conference meeting, a virtual assistant may provide in-meeting assistance responsive to verbalized group intent. The virtual assistant may automatically perform actions, such as recording notes, creating calendar events, and obtaining information derived from previous meetings, in response to the detected group intent. The virtual assistant may also provide pre-meeting assistance.
Claims20
A method for identifying and handling group intent in a video conference using a virtual agent, the method comprising:
- receiving two or more upload audio streams generated at two or more endpoints of a video conferencing system;
- converting the two or more upload audio streams to a plurality of text-user statements;
- determining one or more user-intent names and a corresponding one or more user-context parameters based on one or more of the text-user statements;
- creating a plurality of utterance events based on the plurality of text-user statements, wherein the respective utterance events comprise the respective user-intent names;
- identifying a match between a first utterance event of the plurality of utterance events and a target group intent having a predefined set of required-group-intent attributes, wherein the match is based on the utterance event satisfying a portion of the set of required-group-intent attributes, and wherein satisfying all of the required-group-intent attributes includes identifying a user intent shared by at least two participants in the video conference;
- determining whether a component of a respective event of a first count of previous utterance events from the plurality of utterance events satisfies a first unsatisfied required-group-intent attribute, and parameterizing the first unsatisfied required-group-intent attribute if so;
- determining whether a component of a respective event of a second count of additional utterance events satisfies the first or a second unsatisfied required-group-intent attribute, and parameterizing one or more of the first and second unsatisfied required-group-intent attributes if so;
- formulating a responsive data query concerning the first, second, or a third unsatisfied required-group-intent attribute;
- providing the responsive data query to an integration service stack, wherein the integration service stack provides an application programming interface for a collection of constituent services;
- formulating a responsive agent statement prompt concerning the first, second, third, or a fourth unsatisfied required-group-intent attribute;
- converting the responsive agent statement prompt to audio for playing at one or more of the endpoints;
- determining whether a component of a respective event of a third count of additional utterance events satisfies the first, second, third, or fourth unsatisfied required-group-intent attribute, and parameterizing one or more of the first, second third and fourth unsatisfied required-group-intent attributes if so;
- upon validation of the target group intent, causing a responsive action to be executed, wherein validation is achieved when all of the required-group-intent attributes are satisfied.
10 more elements
A video conferencing system, comprising:
- one or more processors;
- one or more storage devices communicatively coupled to the one or more processors;
- and
- instructions on the one or more storage devices that, when executed by the one or more processors, cause the one or more processors to:
- receive two or more upload audio streams generated at two or more endpoints of a video conferencing system;
- convert the two or more upload audio streams to a plurality of text-user statements;
- determine one or more user-intent names and a corresponding one or more user-context parameters based on one or more of the text-user statements;
- create a plurality of utterance events based on the plurality of text-user statements, wherein the respective utterance events comprise the respective user-intent names;
- identify a match between a first utterance event of the plurality of utterance events and a target group intent having a predefined set of required-group-intent attributes, wherein the match is based on the utterance event satisfying a portion of the set of required-group-intent attributes, and wherein satisfying all of the required-group-intent attributes includes identifying a user intent shared by at least two participants in the video conference;
- determine whether a component of a respective event of a first count of previous utterance events from the plurality of utterance events satisfies a first unsatisfied required-group-intent attribute, and parameterize the first unsatisfied required-group-intent attribute if so;
- determine whether a component of a respective event of a second count of additional utterance events satisfies the first or a second unsatisfied required-group-intent attribute, and parameterize one or more of the first and second unsatisfied required-group-intent attributes if so;
- formulate a responsive data query concerning the first, second, or a third unsatisfied required-group-intent attribute;
- provide the responsive data query to an integration service stack, wherein the integration service stack provides an application programming interface for a collection of constituent services;
- formulate a responsive agent statement prompt concerning the first, second, third, or a fourth unsatisfied required-group-intent attribute;
- convert the responsive agent statement prompt to audio for playing at one or more of the endpoints;
- determine whether a component of a respective event of a third count of additional utterance events satisfies the first, second, third, or fourth unsatisfied required-group-intent attribute, and parameterize one or more of the first, second, third and fourth unsatisfied required-group-intent attributes if so;
- upon validation of the target group intent, cause a responsive action to be executed, wherein validation is achieved when all of the required-group-intent attributes are satisfied.
14 more elements
A non-transitory machine-readable storage medium comprising instructions that, when executed by a processor, cause the processor to:
- receive two or more upload audio streams generated at two or more endpoints of a video conferencing system;
- convert the two or more upload audio streams to a plurality of text-user statements;
- determine one or more user-intent names and a corresponding one or more user-context parameters based on one or more of the text-user statements;
- create a plurality of utterance events based on the plurality of text-user statements, wherein the respective utterance events comprise the respective user-intent names;
- identify a match between a first utterance event of the plurality of utterance events and a target group intent having a predefined set of required-group-intent attributes, wherein the match is based on the utterance event satisfying a portion of the set of required-group-intent attributes, and wherein satisfying all of the required-group-intent attributes includes identifying a user intent shared by at least two participants in the video conference;
- determine whether a component of a respective event of a first count of previous utterance events from the plurality of utterance events satisfies a first unsatisfied required-group-intent attribute, and parameterize the first unsatisfied required-group-intent attribute if so;
- determine whether a component of a respective event of a second count of additional utterance events satisfies the first or a second unsatisfied required-group-intent attribute, and parameterize one or more of the first and second unsatisfied required-group-intent attributes if so;
- formulate a responsive data query concerning the first, second, or a third unsatisfied required-group-intent attribute;
- provide the responsive data query to an integration service stack, wherein the integration service stack provides an application programming interface for a collection of constituent services;
- formulate a responsive agent statement prompt concerning the first, second, third, or a fourth unsatisfied required-group-intent attribute;
- convert the responsive agent statement prompt to audio for playing at one or more of the endpoints;
- determine whether a component of a respective event of a third count of additional utterance events satisfies the first, second, third, or fourth unsatisfied required-group-intent attribute, and parameterize one or more of the first, second, third and fourth unsatisfied required-group-intent attributes if so;
- upon validation of the target group intent, cause a responsive action to be executed, wherein validation is achieved when all of the required-group-intent attributes are satisfied.
10 more elements
Dependencies are inferred from the claim text.