Abstract: Techniques for generating a recommendation for content based on user-device dialogue are described. In an example, a computer system receives a request for audio output at a user device. The computer system determines a response for the request. The computer system determines a topic and value associated with the request and the response. The computer system also determines a recommendation for content based on the request and the response. The computer system generates prompt data including a question of whether to initiate an action associated with the recommendation. The computer system sends the prompt data and the response data to the user device or a different user device.

Abstract: Systems, methods, and non-transitory computer-readable media can be configured to determine first audio associated with a first user and second audio associated with a second user, the first user and the second user associated with a communication session. The second audio can be muted based on a determination that the first audio and the second audio overlap. The second audio can be provided based on completion of the first audio.

Abstract: A crop yield estimation method based on spatio-temporal deep learning including: obtaining regional historical crop yield data and meteorological data, preprocessing the meteorological data and the yield data to respectively obtain meteorological parameters and a detrended yield as input and output of the crop yield spatio-temporal deep learning model; constructing the spatio-temporal deep learning model for crop yield estimation, and optimizing hyperparameters; and building a training set by taking the meteorological parameters as an input and the detrended yield as output to train the model and obtain parameters of the model; for the crop yield to be estimated, feeding meteorological parameters into the trained model, and obtaining the crop yield estimation result. The model combined temporal and spatial learning to achieve better crop yield estimation accuracy and stability at large spatial scales.

Abstract: A method for detection of a keyword in a continuous stream of audio signal, by using a dilated convolutional neural network (DCNN), implemented by one or more computers embedded on a device, the dilated convolutional network (DCNN) comprising a plurality of dilation layers (DL), including an input layer (IL) and an output layer (OL), each layer of the plurality of dilation layers (DL) comprising gated activation units, and skip-connections to the output layer (OL), the dilated convolutional network (DCNN) being configured to generate an output detection signal when a predetermined keyword is present in the continuous stream of audio signal, the generation of the output detection signal being based on a sequence (SSM) of successive measurements (SM) provided to the input layer (IL), each successive measurement (SM) of the sequence (SSM) being measured on a corresponding frame from a sequence of successive frames extracted from the continuous stream of audio signal, at a plurality of successive time steps.

Abstract: Determining which reference-level pressures, from among a plurality of available reference-level pressures, are used when estimating an altitude of a mobile device. Different systems and methods determine isobars based on reference-level pressures of weather stations, and then use the isobars in different ways to identify particular reference-level pressures for use in estimated an altitude of a mobile device. One approach determines the smallest distance between an initial estimated position of a mobile device and a neighboring isobar, and then uses that distance to identify reference-level pressures. Another approach identifies reference-level pressures between an isobar on which an initial estimated position of a mobile device is location and a neighboring isobar. Yet another approach compares the number of identified reference-level pressures and/or locations of the identified reference-level pressures against threshold conditions before determining which reference-level pressures to use.

Abstract: Systems and methods are provided to retrieve or analyze usage data collected from a device or a facility where the device, optionally with devices are located, and identify useful features for making a diagnosis of the device. The diagnosis can be made before a system failure to reduce down time and inefficient use of the device, or after the system failure to expedite and facilitate diagnosis and repair. In addition to the usage data, such as energy and resource consumption, the system can also obtain information relating to the facility and the device's external environment which can be used for normalizing the usage data. Further, based on the diagnosis, the system can make suitable recommendations for repair, replacement, maintenance and upgrade.

Abstract: Systems, apparatuses, methods and computer processes to separate and select frequency components of a signal in real time without phase delay by predicting or forecasting future values of the signal and using that forecast and past measurements in a noncausal zero-phase filtering algorithm are provided. The method of separating and selecting frequency components of a signal in real time without phase delay using a zero-phase filter, comprises obtaining past measurements of the signal; obtaining predicted values of the signal; and using the predicted values of the signal and the past measurements of the signal as components of an input signal in a noncausal zero-phase filtering algorithm for a zero-phase filter, the zero-phase filter producing an output signal.

Abstract: A method, computer system, and a computer program product for context-based wake word threshold modification is provided. The present invention may include receiving an audio data generated by a computing device operating in a sleep mode. The present invention may include processing the received audio data to determine a context associated with a user of the computing device. The present invention may include in response to detecting a potential wake word in the received audio data, calculating a wake word confidence score associated with an accuracy of the detected potential wake word. The present invention may include retrieving a wake word threshold, wherein the wake word threshold is determined based on the context associated with the user of the computing device. The present invention may include, in response to the calculated wake word confidence score exceeding the retrieved wake word threshold, activating a wake mode of the computing device.

Abstract: A computer implemented method may include executing a first simulation test for testing a device under test (DUT) and a corresponding test environment; saving a snapshot image of the DUT and of the corresponding test environment upon completion of initialization actions included in the first simulation test to configure the DUT; compiling a DUT part of a second simulation test into the saved snapshot image of the DUT to obtain a restore image for the DUT; loading the restore image of the DUT and restoring the snapshot image of the test environment; loading a test environment part of the second simulation test; and executing the second simulation test on the DUT and corresponding test environment.

Abstract: A computing system is configured to generate a transformer-transducer-based deep neural network. The transformer-transducer-based deep neural network comprises a transformer encoder network and a transducer predictor network. The transformer encoder network has a plurality of layers, each of which includes a multi-head attention network sublayer and a feed-forward network sublayer. The computing system trains an end-to-end (E2E) automatic speech recognition (ASR) model, using the transformer-transducer-based deep neural network. The E2E ASR model has one or more adjustable hyperparameters that are configured to dynamically adjust an efficiency or a performance of E2E ASR model when the E2E ASR model is deployed onto a device or executed by the device.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining, for each of one or more categorical features, a respective vocabulary of categorical feature values of the categorical feature that should be active during processing of inputs by a machine learning model. In one aspect, a method comprises: generating a batch of output sequences, each output sequence in the batch specifying, for each of the categorical features, a respective vocabulary of categorical feature values of the categorical feature that should be active; for each output sequence in the batch, determining a performance metric of the machine learning model on a machine learning task after the machine learning model has been trained to perform the machine learning task with only the respective vocabulary of categorical feature values of each categorical feature specified by the output sequence being active.

Abstract: An electronic apparatus is provided. The electronic apparatus includes a communication interface and a processor configured to, based on receiving a first audio signal from a first sensor device from among a plurality of sensor devices through the communication interface, perform an operation corresponding to the first audio signal, and control the first sensor device to output a second audio signal indicating a result of performing the operation corresponding to the first audio signal, and based on receiving a third audio signal from a second sensor device from among the plurality of sensor devices within a threshold time from a time of controlling the first sensor device through the communication interface, determine whether to process the third audio signal based on the second audio signal and the third audio signal.

Abstract: A particle measurement system and method of operation thereof are described. The system and method render a characteristic for a set of particles measured while passing through a measurement volume. The system includes a source that generates a particle-laden field containing the set of particles. The system further includes a sensor that generates a raw particle data corresponding to the set particles passing through the measurement volume of the particle measurement system, where the raw particle data comprises a set of raw particle records and each of one of the raw particle records includes a particle data content. A preconditioning stage carries out a preconditioning operation on the particle data content of the set of raw particle records to render a conditioned input data. A machine learning stage processes the conditioned input data to render an output characteristic parameter value for the set of particles.

Abstract: A device for outputting sound and a method therefor are provided. The sound output method includes predicting external sound to be received from an external environment, variably adjusting sound to be output from the device, based on the predicted external sound, and outputting the adjusted sound.

Abstract: Techniques for calibrating listening devices are disclosed herein. The techniques include emitting a predetermined audio signal using an outward-facing transducer located on a first portion of a head-mounted device worn by the user, receiving the predetermined audio signal at a microphone located on a second portion of the head-mounted device, the second portion being different from the first portion, determining a transfer function for the user based on the received predetermined audio signal, and applying the transfer function to audio signals transmitted to the user.

Abstract: A wireless or remote thermometer connected with an artificial intelligence (AI) system. The thermometer may be in communication with a voice-activated AI system to implement operation thereof, such as a cloud-based AI system implemented on a smart audio interface. User-accessible controls for the thermometer may be activated using the voice-activated AI system on the audio interface. The thermometer may include a wireless transceiver for communication with a user. The thermometer collects temperature measurement data to remotely monitor the temperature of food or other materials. The thermometer connects and communicates wirelessly with a receiver unit, such as user smartphone, tablet, or other computerized device. The thermometer unit sends data, alerts or notifications to the delegated receiver, smart device, and/or audio interface, to the user.

Abstract: A device monitors a communication between a user associated with a user device and a service representative associated with a service representative device, and causes a natural language processing model to perform a natural language processing analysis of a user input of the communication to identify a topic associated with the communication. The device determines a first score associated with the topic, and determines a second score associated with enabling the communication, where the first score and second score indicate a service performance score of an entity. The device causes a sentiment analysis model to perform a sentiment analysis of the communication to determine a sentiment score indicating a level of satisfaction the user has relative to the topic. The device updates a transaction protocol associated with the topic based on the service performance score, and/or updates a communication processing protocol associated with the communication based on the sentiment score.

Abstract: An overall performance metric of a computer system may be determined for each bin of the set of analysis bins. In case one or more bins of the set of analysis bins do not have at least a predefined minimum number of records, a new set of analysis bins may be redefined by joining analysis bins of the set of analysis bins. For each bin of the redefined set of bins a machine learning (ML) performance metric of the ML model may be computed. The ML performance metric may be estimated for the set of analysis bins using the ML performance metrics of the redefined bins. The computer system may be configured based on a correlation over the set of analysis bins between the computed overall performance metric and the ML performance metric.

Abstract: Systems and methods are disclosed that provide contextual tracking information to tracking sensor systems to provide accurate and efficient object tracking. Contextual data of a first tracking sensor system is used to identify a tracked object of a second tracking sensor system.

Abstract: A method for training a backpropagation-enabled regression process is used for predicting values of an attribute of subsurface data. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A predicted value of the attribute has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.