Back to results
US 12,764,230 B2Granted
Alert generator for adaptive closed loop communication system
AssigneeGridspace Inc.
InventorsAnthony Scodary, Nicolas Benitez
Life of this patent
Estimating the term…
Abstract
An alert generator in a communication system for processing a call includes at least one machine learning model generating call classifiers from outputs of an audio signal processor and a natural language processor configure to operate on the call. Heuristic logic is configured to transform the call classifiers into a plurality of weighted sub-metrics for the call, and aggregate normalized Gaussian logic is configured to transform the weighted sub-metrics into a metric control. A threshold analyzer is configured to generate an alert signal to the communication system based on the metric control meeting a condition.
Claims20
A communication system for processing a call, the communication system comprising:
- an audio analysis system configured to:
- receive audio content of the call;
- identify speech vocal content from the audio content;
- identify speech semantic content from the audio content;
- determine a combined message content, the combined message content including a transformation of the speech semantic content and the speech vocal content;
- determine a model control structure from the combined message content;
- and
- apply the model control structure as a feedback control to influence creation of future audio records on one or more machines, wherein the model control structure is a grade of the audio content, the one or more machines comprising a machine display, the machine display altered to display the grade;
- and
- an alert generator, the alert generator comprising:
- at least one machine learning model generating call classifiers from outputs of an audio signal processor and a natural language processor configured to operate on the call;
- heuristic logic configured to transform the call classifiers into a plurality of weighted sub-metrics for the call;
- aggregate normalized Gaussian logic to transform the weighted sub-metrics into a metric control;
- and
- a threshold analyzer configured to generate an alert signal to the communication system based on the metric control meeting a condition.
12 more elements
A method for processing a call in a communication system, the method comprising:
- receiving audio content for a call at an audio analysis system;
- identifying speech vocal content from the audio content;
- identifying speech semantic content from the audio content;
- determining a combined message content, the combined message content comprising a transformation of the speech vocal content and the speech semantic content;
- determining a model control structure from the combined message content;
- applying the model control structure as a feedback control to influence creation of future audio records on one or more machines, wherein the model control structure is a grade of the audio content, the one or more machines each comprising a machine display, the machine display altered to display the grade;
- operating on the audio content for the call using an audio signal processor and a natural language processor;
- operating at least one machine learning model on outputs of the audio signal processor and the natural language processor to generate call classifiers;
- operating heuristic logic to transform the call classifiers into a plurality of weighted sub-metrics for the call;
- applying an aggregate normalized Gaussian transform to convert the weighted sub-metrics into a metric control;
- and
- operating a threshold analyzer of an alert generator to generate an alert signal to the communication system based on the metric control meeting a condition.
9 more elements
A computing apparatus processing a call in a communication system, the computing apparatus comprising:
- a processor;
- and
- a memory storing instructions that, when executed by the processor, configure the computing apparatus to:
- receive audio content for a call;
- identify speech vocal content from the audio content;
- identify speech semantic content from the audio content;
- determine a combined message content, the combined message content comprising a transformation of the speech vocal content and the speech semantic content;
- determine a model control structure from the combined message content;
- apply the model control structure as a feedback control to influence creation of future audio records on one or more machines, wherein the model control structure is a grade of the audio content, the one or more machines comprising a machine display, the machine display altered to display the grade;
- operate on the audio content for the call using an audio signal processor and a natural language processor;
- operate at least one machine learning model on outputs of the audio signal processor and the natural language processor to generate call classifiers;
- operate heuristic logic to transform the call classifiers into a plurality of weighted sub-metrics for the call;
- apply an aggregate normalized Gaussian transform to convert the weighted sub-metrics into a metric control;
- and
- operate a threshold analyzer of an alert generator to generate an alert signal to the communication system based on the metric control meeting a condition.
12 more elements
Dependencies are inferred from the claim text.