Abstract: In some examples, a computing system includes a storage device configured to store a front-end neural network and a back-end model; and processing circuitry. The processing circuitry is configured to: receive a test audio data sample; process, by executing the front-end neural network, the test audio data sample to extract one or more embeddings from the front-end neural network; process, by executing the back-end model, the one or more embeddings to determine a likelihood that indicates whether the test audio data sample represents speech by a particular human; and output an indication as to whether the test audio data sample represents genuine speech by the particular human.

Abstract: A computing system is configured to obtain a plurality of media files that each includes speech of one or more speakers. The computing system is further configured to process the plurality of media files to generate indexed data, wherein the indexed data includes a corresponding embedding for each speaker of the one or more speakers identified in the media file and a corresponding one or more keywords identified in the speech in the media file. The computing system is further configured to receive an indication at least one of a selection of a particular speaker from the one or more speakers or a selection of a particular keyword from a plurality of keywords. The computing system is further configured to generate one or more correlations based on the indexed data. The computing system is further configured to output an alert regarding the one or more correlations.

Abstract: In general, the disclosure describes techniques for detecting synthetic speech of a speaker. In an example, a machine learning system may be configured to generate, using a deep learning model trained to distinguish between synthetic speech and authentic speech, reference embeddings for the speaker that characterize a first set of acoustic features and a first set of phonetic features associated with the speaker. The machine learning system may further be configured to generate, using the deep learning model, a test embedding for an audio clip that characterizes a second set of acoustic features and a second set of phonetic features associated with the audio clip. The machine learning system may further be configured to compute a score based on the test embedding and the reference embeddings. The machine learning system may further be configured to output, based on the score, an indication of whether the audio clip includes synthetic speech.

Abstract: The disclosed technologies include methods for generating a calibration model using data that is selected to match the conditions of a particular trial that involves an automated comparison of data samples, such as a comparison-based trial performed by an audio-based recognition, identification, or detection system. The disclosed technologies also include improved methods for selecting candidate data used to build the calibration model. The disclosed technologies further include methods for evaluating the performance of the calibration model and for rejecting a trial when not enough matched candidate data is available to build the calibration model. The disclosed technologies additionally include the use of regularization and automated data generation techniques to further improve the robustness of the calibration model.

Abstract: In an embodiment, the disclosed technologies include automatically recognizing speech content of an audio stream that may contain multiple different classes of speech content, by receiving, by an audio capture device, an audio stream; outputting, by one or more classifiers, in response to an inputting to the one or more classifiers of digital data that has been extracted from the audio stream, score data; where a score of the score data indicates a likelihood that a particular time segment of the audio stream contains speech of a particular class; where the one or more classifiers use one or more machine-learned models that have been trained to recognize audio of one or more particular classes to determine the score data; using a sliding time window process, selecting particular scores from the score data; using the selected particular scores, determining and outputting one or more decisions as to whether one or more particular time segments of the audio stream contain speech of one or more particular classes

Abstract: In an embodiment, the disclosed technologies include automatically recognizing speech content of an audio stream that may contain multiple different classes of speech content, by receiving, by an audio capture device, an audio stream; outputting, by one or more classifiers, in response to an inputting to the one or more classifiers of digital data that has been extracted from the audio stream, score data; where a score of the score data indicates a likelihood that a particular time segment of the audio stream contains speech of a particular class; where the one or more classifiers use one or more machine-learned models that have been trained to recognize audio of one or more particular classes to determine the score data; using a sliding time window process, selecting particular scores from the score data; using the selected particular scores, determining and outputting one or more decisions as to whether one or more particular time segments of the audio stream contain speech of one or more particular classes

Abstract: A spoken command analyzer computing system includes technologies configured to analyze information extracted from a speech sample and, using a joint speaker and phonetic content model, both determine whether the analyzed speech includes certain content (e.g., a command) and to identify the identity of the human speaker of the speech. In response to determining that the identity matches the authorized user's identity and determining that the analyzed speech includes the modeled content (e.g., command), an action corresponding to the verified content (e.g., command) is performed by an associated device.

Abstract: The disclosed technologies include methods for generating a calibration model using data that is selected to match the conditions of a particular trial that involves an automated comparison of data samples, such as a comparison-based trial performed by an audio-based recognition, identification, or detection system. The disclosed technologies also include improved methods for selecting candidate data used to build the calibration model. The disclosed technologies further include methods for evaluating the performance of the calibration model and for rejecting a trial when not enough matched candidate data is available to build the calibration model. The disclosed technologies additionally include the use of regularization and automated data generation techniques to further improve the robustness of the calibration model.

Abstract: An audio file analyzer computing system includes technologies to, among other things, localize audio events of interest (such as speakers of interest) within an audio file that includes multiple different classes (e.g., different speakers) of audio. The illustrative audio file analyzer computing system uses a seed segment to perform a semi-supervised diarization of the audio file. The seed segment is pre-selected, such as by a human person using an interactive graphical user interface.

Abstract: An audio file analyzer computing system includes technologies to, among other things, localize audio events of interest (such as speakers of interest) within an audio file that includes multiple different classes (e.g., different speakers) of audio. The illustrative audio file analyzer computing system uses a seed segment to perform a semi-supervised diarization of the audio file. The seed segment is pre-selected, such as by a human person using an interactive graphical user interface.

Abstract: A spoken command analyzer computing system includes technologies configured to analyze information extracted from a speech sample and, using a joint speaker and phonetic content model, both determine whether the analyzed speech includes certain content (e.g., a command) and to identify the identity of the human speaker of the speech. In response to determining that the identity matches the authorized user's identity and determining that the analyzed speech includes the modeled content (e.g., command), an action corresponding to the verified content (e.g., command) is performed by an associated device.