Abstract: A system for generating channel-compensated features of a speech signal includes a channel noise simulator that degrades the speech signal, a feed forward convolutional neural network (CNN) that generates channel-compensated features of the degraded speech signal, and a loss function that computes a difference between the channel-compensated features and handcrafted features for the same raw speech signal. Each loss result may be used to update connection weights of the CNN until a predetermined threshold loss is satisfied, and the CNN may be used as a front-end for a deep neural network (DNN) for speaker recognition/verification. The DNN may include convolutional layers, a bottleneck features layer, multiple fully-connected layers and an output layer. The bottleneck features may be used to update connection weights of the convolutional layers, and dropout may be applied to the convolutional layers.

Abstract: Systems and methods for distributed voice processing are disclosed herein. In one example, the method includes detecting sound via a microphone array of a first playback device and analyzing, via a first wake-word engine of the first playback device, the detected sound. The first playback device may transmit data associated with the detected sound to a second playback device over a local area network. A second wake-word engine of the second playback device may analyze the transmitted data associated with the detected sound. The method may further include identifying that the detected sound contains either a first wake word or a second wake word based on the analysis via the first and second wake-word engines, respectively. Based on the identification, sound data corresponding to the detected sound may be transmitted over a wide area network to a remote computing device associated with a particular voice assistant service.

Abstract: Devices and techniques are generally described for a speech processing routing architecture. First input data representing an input request may be received. First data may be sent to a first skill representing a first request for the first skill to evaluate an ability of the first skill to process the first input data. Second data may be sent to a second skill representing a second request for the second skill to evaluate an ability of the second skill to process the first input data. Third data may be received from the first skill indicating a first action performed by the first skill in response to receipt of the first input data. Fourth data may be received from the second skill indicating a second action performed by the second skill. The first skill may be selected for processing the first input data.

Abstract: A device includes a memory adapted to store a list in a file or database comprising a plurality of vocabulary words in a first language and, for each vocabulary word, a corresponding word in a second language, a display device, and a processor. The processor is adapted to receive a plurality of words in the first language, select one or more words among the plurality of words, based on one or more predetermined criteria, translate, match or equate the one or more selected words from the first language to words of the second language, and cause the display device to display the plurality of words, wherein one or more first words that are in the plurality of words and are not among the one or more selected words which are displayed in the first language and one or more second words that are in the plurality of words and are among the one or more selected words are displayed in the second language.

Abstract: Devices and techniques are generally described for inference reduction in natural language processing using semantic similarity-based caching. In various examples, first automatic speech recognition (ASR) data representing a first natural language input may be determined. A cache may be searched using the first ASR data. A first skill associated with the first ASR data may be determined from the cache. In some examples, first intent data representing a semantic interpretation of the first natural language input data may be determined by using a first natural language process associated with the first skill.

Abstract: A User Equipment (UE) is operative to generate CN (Comfort Noise) control parameters, e.g., as part of audio-decoding processing by the UE. A buffer of a predetermined size implemented in the UE is configured to store CN parameters for SID (Silence Insertion Descriptor) frames and active hangover frames. Processing circuitry of the UE is configured to determine a CN parameter subset relevant for SID frames based on the age of the stored CN parameters and on residual energies, and use the determined CN parameter subset to determine CN control parameters for a first SID frame following an active signal frame.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable medium for data input with multi-format validation. The method may include receiving data input via a microphone mounted on a user device and receiving the data input via a camera mounted on the user device. Additionally, the method may include comparing the data input via the microphone and the data input via the camera and determining whether the comparison of the data input exceeds a predetermined confidence level. Additionally, the method may include storing the data input, upon determining that the comparison of the data input exceeds the predetermined confidence level and presenting to the user a notification of validation upon determining that the comparison of the data input does not exceed the predetermined confidence level. Additionally, the method may include receiving from the user a validation of the data input based on the notification of validation and storing the data input based on the validation of the data input.

Abstract: An information processing device includes a network interface and a processor. The processor is configured to: acquire voice data via the network interface, analyze the acquired voice data, based on a result of the analysis, determine a search condition including one or more keywords for searching for one or more items, perform a search using the determined search condition, generate a first text indicating an item found by the search, and controls the network interface to output the generated first text. The processor is further configured to, when two or more items are found by the search, generate a second text suggesting another keyword other than said one or more keywords that have been used for the search, and controls the network interface to output the generated second text.

Abstract: A spectrum coding method includes quantizing spectral data of a current band based on a first quantization scheme, generating a lower bit of the current band using the spectral data and the quantized spectral data, quantizing a sequence of lower bits including the lower bit of the current band based on a second quantization scheme, and generating a bitstream based on a upper bit excluding N bits, where N is 1 or greater, from the quantized spectral data and the quantized sequence of lower bits.

Abstract: This disclosure describes, in part, techniques and systems for generating and outputting immersive, multi-device content items in user environment, such as connected homes, offices, and the like. For example, the techniques and systems may output different portions of content on different devices within a user environment based on information such as respective capabilities of the devices, a current location of the user within the environment, a time of day, which user(s) are present in the environment, and/or the like.

Abstract: A speech-processing system may provide access to multiple virtual assistants. Speech-processing systems may perform actions for or on behalf of users with the aid of skills; e.g., a shopping skill, navigation skill, communications skill, etc. Some skills may be associated with more than one assistant. The speech-processing system may determine which assistant to invoke upon receiving a command from a user device. The identity of the virtual assistant is propagated to the skill and the device, as well as other components of the speech-processing system. In some cases, however, a multi-assistant skill may determine that an assistant other than the one initially selected by the speech-processing system is to handle the command. The skill may send the identity of the new assistant back to the speech-processing system. The speech-processing system may restart the command dissemination process to provide each component of the system with the updated assistant identity.

Abstract: A computer-implemented method for facilitating navigation of an oil-gas domain application using a virtual assistant integrated within the oil-gas domain application includes generating a trained model for responding to utterances received from a user via a virtual assistant integrated within an oil-gas domain application. The trained model links the utterances to respective actions and responses; receiving a user utterance via the virtual assistant integrated within the oil-gas domain application. The method further includes determining a response to the user utterance using the trained model, wherein the response is associated with performing an action within the oil-gas domain application; and providing the response to the virtual assistant to cause the virtual assistant to execute the action within the oil-gas domain application.

Abstract: Systems and methods are provided to determine consensus values for duplicate fields in a document or form.

Abstract: A computer server system comprises a communications module; a processor coupled with the communications module; and a memory coupled to the processor and storing processor-executable instructions which, when executed by the processor, configure the processor to receive, via the communications module and from a server associated with a first device, a request to perform an operation; determine that the first device cannot perform the operation; send, via the communications module and to the server associated with the first device, a signal causing the first device to output a message indicating that the first device cannot perform the operation and requesting authentication from a second device; receive, via the communications module and from the second device, a signal including authentication information; and send, via the communications module and to the second device, a signal including a selectable option to perform the operation.

Abstract: Speech fluency evaluation and feedback tools are described. A computing device such as a smartphone may be used to collect speech (and/or other data). The collected data may be analyzed to detect various speech events (e.g., stuttering) and feedback may be generated and provided based on the detected speech events. The collected data may be used to generate a fluency score or other performance metric associated with speech. Collected data may be provided to a practitioner such as a speech therapist or physician for improved analysis and/or treatment.

Abstract: A method for exiting a voice skill, an apparatus, a device, and a storage medium are provided by embodiments of the present disclosure, wherein a user voice instruction is received; a target exit intention corresponding to the user voice instruction is identified according to the user voice instruction and a grammar rule of a preset exit intention; and a corresponding operation is executed on a current voice skill of a device according to the target exit intention. The embodiments of the present disclosure refine and expand the user's exit intention. After the target exit intention to which the user voice instruction belongs is identified, the corresponding operation is executed according to the target exit intention so as to meet the users' different exit requirements for the voice skills, enhance the fluency and convenience of user interaction with the device and improve the user's exit experience when using the voice skills.

Abstract: A method for encoding a high frequency signal includes determining a signal type of a high frequency signal of a current frame, smoothing and scaling time envelopes of the high frequency signal of the current frame and obtaining time envelopes of the high frequency signal of the current frame that require to be encoded when the high frequency signal of the current frame is a non-transient signal and a high frequency signal of the previous frame is a transient signal, and quantizing and encoding the time envelopes of the high frequency signal of the current frame that require to be encoded, and frequency information and signal type information of the high frequency signal of the current frame.

Abstract: Aspects relate to apparatuses and methods for selectively inserting text into a video resume. An exemplary apparatus includes a processor and a memory communicatively connected to the processor, the memory containing instructions configuring the processor to receive a video resume from a user, divide the video resume is into temporal sections, acquire a plurality of textual inputs from a user, wherein the plurality of textual inputs pertains to the same user of received video resume, classify the plurality of textual inputs to corresponding temporal sections of the received video resume and display, as a function of the classification, the received video resume with a corresponding plurality of textual inputs.

Abstract: A robot according to the present disclosure comprises: a microphone; a camera disposed to face a predetermined direction; and a processor configured to: inactivate driving of the camera and activate driving of the microphone, if a driving mode of the robot is set to a user monitoring mode; acquire a sound signal through the microphone; activate the driving of the camera based on an event estimated from the acquired sound signal; confirm the event from the image acquired through the camera; and control at least one constituent included in the robot to perform an operation based on the confirmed event.

Abstract: A voice inputting device inputs a voice operation of a user, and transmits voice data based on the voice operation to a first cloud server. The first cloud server receives the voice data from the voice inputting device, analyzes the received voice data, and determines an operational skill level of the user and the details of the voice operation. A second cloud server generates a control command for an air conditioner based on the operational skill level and the details of the voice operation determined by the first cloud server, and transmits the generated control command to the air conditioner.