Abstract: Implementations disclosed herein are directed to federated learning of machine learning (“ML”) model(s) based on gradient(s) generated at corresponding client devices and a remote system. Processor(s) of the corresponding client devices can process client data generated locally at the corresponding client devices using corresponding on-device ML model(s) to generate corresponding predicted outputs, generate corresponding client gradients based on the corresponding predicted outputs, and transmit the corresponding client gradients to the remote system. Processor(s) of the remote system can process remote data obtained from remote database(s) using global ML model(s) to generate additional corresponding predicted outputs, generate corresponding remote gradients based on the additional corresponding predicted outputs. Further, the remote system can utilize the corresponding client gradients and the corresponding remote gradients to update the global ML model(s) or weights thereof.

Abstract: Disclosed embodiments may include a method that includes receiving description data in an originating language for a user and a location associated with the user, identifying one or more names from the description data, retrieving additional data in the originating language based on the one or more names, generating enhanced description data in the originating language for the user based on the description data and the additional data, identifying a target language based on the location associated with the user, selecting a first trained neural network from a plurality of trained neural networks based the target language, providing the enhanced description data in the originating language to the first trained neural network, translating, via the first trained neural network, the enhanced description data from the originating language to the target language, and generating a graphical user interface for display that comprises the enhanced description data in the target language.

Abstract: Methods and systems are disclosed configured to perform automated volume leveling on speech content in an audio file containing speech and non-speech segments. A low pass filter and a high pass filter may be applied to the audio data, and normalization may be performed. Speech and non-speech segments may be detected. Gain adjustments may be made to achieve a substantially constant short term loudness. Processing may be applied to enhance speech parameters, such as attack and release. An upward expander may be used to achieve a target loudness level. A limited and/or dynamic range compressor may be utilized to satisfy true peak and/or short term loudness specifications. A file of processed audio data may be generated and transmitted to one or more destinations for broadcast and/or streaming.

Abstract: An autonomous agent receives a user query comprising the complex question. The agent can obtain, from a corpus of unstructured texts, an answer candidate text corresponding to the user query and comprising text from which the answer is subsequently identified. The agent may generate first linguistic data corresponding to the user query and second linguistic data corresponding to the answer candidate text. Each instance of linguistic data may comprise a combination of respective syntactic data, semantic data, and discourse data generated from the user query and/or answer candidate text. Both instances of linguistic data may be provided to a machine-learning model that has been previously trained to output an answer identified from an instance of unstructured text (e.g., the answer candidate text). The model may output the answer identified from the answer candidate text, which in turn may be provided in response to the user query.

Abstract: A translating method using visually represented elements, and device therefor is provided.

Abstract: The present invention relates to a method for assisting an interviewing party in deciding a response action in response to an interview between said interviewing party and an interviewee party. The method comprises providing a processing unit and inputting the voice of the interviewee party into the processing unit as an electronic signal, and processing the electronic signal by means of said processing unit in parallel with the interview taking place. The method further includes an anomaly routine comprising a statistically learned model, and by means of said statistically learned model determining a respective number of samples of said sequence of samples being an anomaly of said statistically learned model and returning to said anomaly routine for processing a subsequent number of samples of said sequence of samples by said anomaly routine.

Abstract: A system for condensing user communications relating to a topic is provided. The system may include a processor and a non-transitory memory. The processor may: designate a topic of user interest; retrieve legacy communications; and remove duplicative communications. The processor may form a topic-centric training set for a neural network. The topic-centric training set may be based on the legacy communications, legacy intelligence, and the plurality of outcomes and may be delimited by an analysis of the database. The processor may synthesize the neural network using the topic-centric training set in order to assign individual weights to each of a plurality of nodes in the neural network. In response to a selection of the topic of user interest, the processor may generate a plurality of user options based on the neural network. The system may include a display in order to prompt the user to select one of the options.

Abstract: Disclosed is a dialogue emotion correction method based on a graph neural network, including: extracting acoustic features, text features, and image features from a video file to fuse them into multi-modal features; obtaining an emotion prediction result of each sentence of a dialogue in the video file by using the multi-modal features; fusing the emotion prediction result of each sentence with interaction information between talkers in the video file to obtain interaction information fused emotion features; combining, on the basis of the interaction information fused emotion features, with context-dependence relationship in the dialogue to obtain time-series information fused emotion features; correcting, by using the time-series information fused emotion features, the emotion prediction result of each sentence that is obtained previously as to obtain a more accurate emotion recognition result.

Abstract: A system for condensing user communications relating to a topic is provided. The system may include a processor and a non-transitory memory. The processor may: designate a topic of user interest; retrieve legacy communications; and remove duplicative communications. The processor may form a topic-centric training set for a neural network. The topic-centric training set may be based on the legacy communications, legacy intelligence, and the plurality of outcomes and may be delimited by an analysis of the database. The processor may synthesize the neural network using the topic-centric training set in order to assign individual weights to each of a plurality of nodes in the neural network. In response to a selection of the topic of user interest, the processor may generate a plurality of user options based on the neural network. The system may include a display in order to prompt the user to select one of the options.

Abstract: A method and system for automated voice casting compares candidate voices samples from candidate speakers in a target language with a primary voice sample from a primary speaker in a primary language. Utterances in the audio samples of the candidates speakers and the primary speaker are identified and typed and voice samples generated that meet applicable utterance type criteria. A neural network is used to generate an embedding for the voice samples. A voice sample can include groups of different utterance types and embeddings generated for each utterance group in the voice sample and then combined in a weighted form wherein the resulting embedding emphasizes selected utterance types. Similarities between embeddings for the candidate voice samples relative to the primary voice sample are evaluated and used to select a candidate speaker that is a vocal match.

Abstract: A building management system (BMS) for a medical facility that includes a plurality of rooms with at least one of the rooms having a plurality of sensors. An elevated infection risk determination system is operatively coupled to the plurality of sensors for determining when an elevated infection risk occurs in one or more of the rooms. The BMS may include a memory for storing one or more programmable infection risk compliance parameters, an input port for receiving an elevated infection risk alert for the particular room in the medical facility, a control port for providing control commands to one or more building components of the building management system, and a controller. The controller may be configured to provide control commands via the control port in response to receiving the elevated infection risk alert for the particular room based at least in part on one or more programmable infection risk compliance parameters to help mitigate the elevated infection risk in the particular room.

Abstract: A method collects data in a network by operation of a local sensor of a consumption meter where the network is part of a supply network. The network distributes a consumable good. The sensor has a measuring element providing raw measurement data corresponding to a physical or physicochemical value or parameter. The sensor has a wired and/or radio communication device and a memory. For the determination of the measurement resolution of the sensor the conditions for generating time stampings using a correlation model are determined in advance, on the basis of the correlation model time stampings of successive raw measurement data in the sensor are generated. The time stampings are transmitted over a wired connection and/or a radio link so that on the basis of the time stampings using the correlation model the raw measurement data collected by the measuring element are reconstructed and evaluated and used for network monitoring.

Abstract: A method of attaching to a wireless communication network to receive wireless communication service. The method comprises calculating a checksum by a monitor application over at least a portion of an operating system stored in a wireless communication device, wherein the monitor application is stored in a trusted portion of memory of the device and executes in a trusted execution environment (TEE) of the device and wherein the operating system is authorized by a wireless communication network, comparing the calculated checksum value by the monitor application to a checksum value stored in the TEE, based on determining that the calculated checksum value matches the stored checksum value, rebooting the device to execute the operating system, and transmitting radio access credentials to the wireless communication network by the device, whereby the device authenticates into the wireless coverage of the wireless communication network based on the radio access credentials.

Abstract: Embodiments herein are directed a vehicle authorization system for a parking facility. The system includes a gate assembly and a communication system. The gate assembly includes a gate, a gate actuator configured to move the gate, and a gate controller that receives and transmits wireless long range signals. The gate controller is configured to actuate the gate actuator to move the gate. The communication system includes a remote server computing device that is positioned offsite of the parking facility and a gateway device positioned at the parking facility and communicatively coupled to the gate controller and to the remote server computing device. The remote server computing device provides the wireless long range signals to the gate controller via the gateway device such that a construction material of the parking facility and the housing of the gate assembly minimally interferes with the wireless long range signals.

Abstract: A communications hub for a hydrant can include a plug comprising a non-metallic material and configured to be sealably received within a plug bore defined in a flange of a bonnet of the hydrant, the flange separating an interior cavity of the hydrant from a bonnet cavity of the hydrant; a PCB; at least one battery in electrical communication with the PCB; and an antenna configured for wireless communication with a sensing device located within the interior cavity of the hydrant and in electrical communication with the PCB, the antenna able to receive a wireless signal from the sensing device.

Abstract: A system combining readers for RFID transponders and Bluetooth® devices for for automated roadway tolling and monitoring is described. Because of the ubiquity of Bluetooth® devices, filtering methods are described to narrow received Bluetooth® signals to a particular monitored traffic lane and to further associate the Bluetooth® signal with a registered RFID toll tag in a vehicle. Signal strength of the Bluetooth® signal at each of a plurality of receiver antennas may be used to identify the location of the Bluetooth® device that emitted the signal.

Abstract: A vehicle includes a plurality of cameras configured to capture a surrounding situation of the vehicle, an external speaker configured to output a sound to the outside of the vehicle, a communicator configured to communicate with a first server, and a controller configured to determine a degree of risk of the dangerous situation based on the degree of risk based on a case where a dangerous situation is captured through the plurality of cameras, to control an Ultra Wide Band (UWB) module to detect obstacles around the vehicle, to control at least one of a digital key and an external speaker to output a message for the dangerous situation or an image in which the dangerous situation is captured, and to transmit the image in which the dangerous situation is captured to the first server through the communicator.

Abstract: A system for monitoring compressed fluid sources of an evacuation assembly may comprise a controller and a tangible, non-transitory memory configured to communicate with the controller. The controller may perform operations, which may comprise receiving a series of first pressure measurements from a first sensor system, receiving a series of second pressure measurements from a second sensor system, calculating a first pressure trend using the series of first pressure measurements, calculating a second pressure trend using the series of second pressure measurements, and comparing each of the first pressure trend and the second pressure trend to a threshold pressure trend.

Abstract: Communication networks, relay nodes for communication networks, and methods of transmitting data among a plurality of relay nodes utilizing non-dispersive guided acoustic waves are disclosed herein. The communication networks include an elongate tubular body and a wireless data transmission network including a plurality of relay nodes. The relay nodes include an electro-acoustic transmitter array and an electro-acoustic receiver. The methods include inducing a first acoustic wave within the elongate tubular body with a first relay node, conveying the first acoustic wave along the elongate tubular body to a second relay node, and receiving the first acoustic wave with the second relay node. The methods further include inducing a second acoustic wave within the elongate tubular body with the second relay node, conveying the second acoustic wave along the elongate tubular body to a third relay node, and receiving the second acoustic wave with the third relay node.

Abstract: A collision prevention system for a surgical suite includes a supply unit adjustable along a movement path from an initial position to a subsequent position within the surgical suite. A sensor is operably coupled to the supply unit. The sensor is configured to sense an object proximate to the supply unit. A controller is communicatively coupled with the sensor. The controller monitors the object sensed by the sensor. An alert device is communicatively coupled with the controller. The alert device provides a notification when the object is sensed within the movement path and provides a prevention indicator corresponding to an alternate path.