Abstract: Systems and techniques are generally described for anomaly detection in text. In some examples, text data comprising a plurality of words may be received. An image of a first word of the plurality of words may be generated. A feature representation of the first word may be generated using a variational autoencoder. A score may be generated based at least in part on the feature representation. In various examples, the score may indicate a likelihood that an appearance of the first word in the image of the first word is anomalous with respect to at least some other words of the plurality of words.

Abstract: Systems and methods are described for natural language processing of a text sequence. The system can identify a set of text and location information for the set of text in an image. The set of text may correspond to an input sequence space. The system can project embeddings of the text into a latent space for processing. Further, the system can reproject the processed embeddings from the latent space to the input sequence space. The system may perform multiple stages of projecting the embeddings to the latent space and reprojecting the processed embeddings from the latent space to the input sequence space. The system can route the reprojected embeddings to a neural network that can identify class predictions for elements of the set of text.

Abstract: An in-vehicle infotainment (IVI) system is described that includes a display device, an input device communicatively coupled to the display device, and one or more sensors disposed proximate to the input device, wherein the IVI system is configured to determine whether a driver or a passenger is utilizing the input device based on a detection by the one or more sensors.

Abstract: A device, system, and method perform lossy compression for images and signals by identifying regions with a low density of features. The method performed at a sensor communicatively connected to a receiver includes capturing sensor data. The method includes selecting a position in the sensor data. The method includes determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values. When the local entropy is below a predetermined threshold, the method includes applying a first pre-processing operation to the position that averages features included in the position. The method includes transmitting the pre-processed sensor data corresponding to the position to the receiver.

Abstract: A device, system, and method perform lossy compression for images and signals by identifying regions with a low density of features. The method performed at a sensor communicatively connected to a receiver includes capturing sensor data. The method includes selecting a position in the sensor data. The method includes determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values. When the local entropy is below a predetermined threshold, the method includes applying a first pre-processing operation to the position that averages features included in the position. The method includes transmitting the pre-processed sensor data corresponding to the position to the receiver.

Abstract: A device, system, and method determine a topic based publish and parametric subscribe pattern. The method performed at a gateway of a communications network includes receiving a subscriber request, the subscriber request including a plurality of terms, the terms associated with a topic keyword of the subscriber request, a combination of the terms associated with a parameter of the subscriber request. The method includes receiving a published content item that corresponds to the topic keyword, the published content item including information used to address the request based on the parameter. The method includes generating a further published content item including a result based on the topic keyword, the parameter, and the information.