Abstract: A method of adapting a computerized assistant program to satisfy an updated constraint. The method comprises maintaining a dialogue history including a first utterance that indicates an initial constraint. The method further comprises receiving a second utterance indicating a new constraint that conflicts with the initial constraint. The method further comprises recognizing a revision function statement parametrized by a reference to an initial computerized assistant program configured to satisfy the initial constraint, and a reference to the new constraint. The method further comprises executing instructions derived from the revision function statement to return a revised computerized assistant program that is configured to satisfy the new constraint.

Abstract: Systems and methods are provided for determining hint words that improve the accuracy of automated speech recognition (ASR) systems. Hint words are typically determined in the context of a user issuing voice commands in connection with a voice interface system, however, a voice interface system may capture terms from overheard content and/or conversations. A system may determine a sliding window of hint words using set of qualifier rules. The system may capture audio, e.g., from a conversation or played back content, as a first input and decipher a plurality of words including a qualifying first term added to the hint words. The voice interface system may capture more audio as a second input and decipher a second plurality of words including a qualifying second term. The first term may be removed from the set of hint words, e.g., when the second term is added or after an expiration time.

Abstract: Systems and methods for converting a primary one-dimensional signal into a secondary one-dimensional signal of another modality. The primary signal is spliced into a plurality of consecutive frames. A first linear transformation transforms the frames into corresponding vectors. Positional encodings are provided on the vectors to encode relative positional information associated with each sample within each frame. A multi-head self-attention machine-learning model compares relative importance of the samples within each vector to each other in that vector to yield high-level representation vectors. A second linear transformation transforms the high-level representation vectors into corresponding secondary signal frames. The secondary signal frames are concatenated into a reconstructed one-dimensional secondary signal having a different modality than the primary signal.

Abstract: Methods and systems for propagating a stopping condition through a multiple-producer, multiple-consumer distributed system. The method includes determining the number of active processes in a process layer, determining that a stopping condition is satisfied, generating a sentinel in a source queue, receiving a processing task at a process, determining whether the processing task is a sentinel, terminating the first process, decrementing the number of active processes by one, and generating the sentinel in a destination queue.

Abstract: Described herein are exemplary devices, apparatuses, systems, methods, and non-transitory storage media for providing an application framework. The application framework can provide various machine-learning models to perform a variety of analysis tasks to analyze enterprise data such as communications between one or more employees of an organization and one or more clients of the organization and provide intelligence and insights for a user in the organization. The insights and intelligence can include a recommendation or an observation related to a client or customer of the organization. The recommendation or observation can be provided, for example, in a communication platform, a chatbot, or a variety of other interfaces. Advantageously, to perform an analysis task, the application framework automatically provides to the machine-learning model(s) information in accordance with the enterprise's data sharing and access control requirements to prevent inappropriate access and use of sensitive information.

Abstract: Systems and methods are provided for extracting entities from received speech. The systems and methods perform operations comprising receiving an audio file comprising speech input and processing, by a speech recognition engine, the audio file comprising the speech input to generate an initial character-based representation of the speech input. The operations further comprise processing, by an entity extractor, the initial character-based representation of the speech input to generate an estimated set of entities of the speech input. The operations further comprise generating, by the speech recognition engine, a textual representation of the speech input based on the estimated set of entities of the speech input.

Abstract: Certain examples described herein provide a system for classification of unstructured text data relating to a legal query. The system has a session interface to receive session data relating to the legal query, a text interface to receive unstructured text data from a user, a text pre-processor to apply one or more text pre-processing functions to the unstructured text data to output a structured numeric representation of the unstructured text data, at least one machine learning classifier to map the structured numeric representation of the unstructured text data to one or more classes within a defined set of classes, and a classifier optimizer to process the session data to generate configuration data for the at least one machine learning classifier, the configuration data indicating a subset of the defined set of classes that are valid given the session data.

Abstract: Data is received that encapsulates a video of a subject performing a task. This video is used to generate a transcript using an automatic speech recognition (ASR) system. A plurality of text segments are generated from the transcript and then tokenized. A textual representation of each segment is extracted by a transformer model using the tokenized text segment (i.e., the tokens corresponding to the text segment). Thereafter, for each segment, a fused representation derived from the textual representations and corresponding visual and audio features from the video is generated. A sparse attention machine learning model then selects an optimal slice of the video based on the fused representations. The optimal slice can then be input into one or more machine learning models trained to characterize performance of the task by the subject.

Abstract: Systems and methods detecting a spoken sentence in a speech recognition system are disclosed herein. Speech data is buffered based on an audio signal captured at a computing device operating in an active mode. The speech data is buffered irrespective of whether the speech data comprises a signature word. The buffered speech data is processed to detect a presence of the sentence comprising at least one command and a query for the computing device. Processing the buffered speech data includes detecting the signature word in the buffered speech data, and in response to detecting the signature word in the speech data, initiating detection of the sentence in the buffered speech data.

Abstract: In one aspect, a user's intent can be inferred based on voice analysis during a communications session, and prompts can be presented, or other actions taken, at least partly in response to the inferred intent. For example, a network microphone device (NMD) having one or more microphones can capture voice input and transmit the voice input to remote computing device(s) for a communication session (e.g., a videoconference). The NMD can analyze the voice input to detect one or more utterances. Based on the utterance(s), the NMD can cause a user prompt to be displayed via a display device communicatively coupled to the NMD. The particular prompt can depend at least in part on one or more context parameters associated with the communication session (e.g., a microphone state of one or more users, a screen share state of one or more users, or a recording status of the session, etc.).

Abstract: An audio firewall system has a microphone that generates audio data. A speech-to-text engine converts the audio data to text data. The text data is parsed for a service wake word and corresponding content data. The service wake word identifies one of a local security system and a remote assistant server. A text-to-speech engine converts the service wake word and the corresponding content data to converted audio data. The converted audio data is provided to the remote assistant server. The content data is provided to the local security system. The audio firewall system receives a response from the remote assistant server or the local security system and outputs an audio signal corresponding to the response.

Abstract: An AI control device, which identifies individual users from a plurality of users to receive input data, and is connectable to a server device that generates a trained model based on input data for each user, includes a control unit, and a communication unit connected to the server device. The control unit acquires input data, associates acquired input data and identifying information used to identify the user of the AI control device, and sends the data and information to the server device via the communication unit. The control unit uses the sent acquired input data to execute a trained model that is generated separately from trained models of other users by the server device, and that learns characteristics of acquired input data and detects input data having the same characteristics from unknown input data.

Abstract: Methods and systems for using generative content to improve the ability of an individual to communicate using electronic-assisted communication.

Abstract: A device, comprises a microphone and a communication interface configured to send to a premises automation control core audio input received via the microphone. The premises automation control core is configured to determine, based at least in part on the audio input, a premises automation context associated with the device and send to the device via the communication interface a context data indicating the determined context. The devices also comprises a visual display device, a physical input device, and a processor coupled to the visual display device and the communication interface. The processor is configured to receive the context data and cause the visual display device to provide a visual display associated with the context data.

Abstract: Determining whether, upon cessation of a second automated assistant session that interrupted and supplanted a prior first automated assistant session: (1) to automatically resume the prior first automated assistant session, or (2) to transition to an alternative automated assistant state in which the prior first session is not automatically resumed. Implementations further relate to selectively causing, based on the determining and upon cessation of the second automated assistant session, either the automatic resumption of the prior first automated assistant session that was interrupted, or the transition to the state in which the first session is not automatically resumed.

Abstract: A data processing system implements a method for training machine learning modes, including receiving a set of one or more unlabeled documents associated one or more first categories of documents to be used to train machine learning models to analyze the one or more unlabeled documents, and fine-tuning a first machine learning model and a second machine learning model based on the one or more unlabeled document to enable the first machine learning model to determine a semantic representation of the one or more first categories of document, and to enable the second machine learning model to classify the semantic representations according to the one or more first categories of documents, the first machine learning model and the second machine learning model having been trained using first unlabeled training data including a second plurality of categories of documents that do not include the one or more first categories of documents.

Abstract: Examples relate to a surgical microscope system and to a corresponding system, method and computer program for a surgical microscope system. The system comprises one or more processors and one or more storage devices. The system is configured to obtain an audio signal from a microphone of the surgical microscope system. The system is configured to analyze the audio signal locally to detect one or more spoken commands within the audio signal. A user-specific voice profile is used to determine whether the one or more spoken commands are uttered by a user associated with the user-specific voice profile. The system is configured to control the surgical microscope system based on the detected one or more spoken commands if the one or more spoken commands are uttered by the user associated with the user-specific voice profile.

Abstract: An information processor including: an operation control unit that controls a motion of an autonomous mobile body acting on the basis of recognition processing, in a case where a target sound that is a target voice for voice recognition processing is detected, the operation control unit moving the autonomous mobile body to a position, around an approach target, where an input level of a non-target sound that is not the target voice becomes lower, the approach target being determined on the basis of the target sound.

Abstract: A method includes receiving a spoken utterance that includes a plurality of words, and generating, using a neural network-based utterance classifier comprising a stack of multiple Long-Short Term Memory (LSTM) layers, a respective textual representation for each word of the of the plurality of words of the spoken utterance. The neural network-based utterance classifier trained on negative training examples of spoken utterances not directed toward an automated assistant server. The method further including determining, using the respective textual representation generated for each word of the plurality of words of the spoken utterance, that the spoken utterance is one of directed toward the automated assistant server or not directed toward the automated assistant server, and when the spoken utterance is directed toward the automated assistant server, generating instructions that cause the automated assistant server to generate a response to the spoken utterance.

Abstract: Systems and methods for audio processing are described. An audio processing system receives audio content that includes a voice sample. The audio processing system analyzes the voice sample to identify a sound type in the voice sample. The sound type corresponds to pronunciation of at least one specified character in the voice sample. The audio processing system generates a filtered voice sample at least in part by filtering the voice sample to modify the sound type. The audio processing system outputs the filtered voice sample.