Abstract: A vehicle control apparatus to be used in a vehicle controllable on the basis of a voice input includes a determination unit and an input unit. The determination unit is configured to determine whether a main operator of the vehicle is in a predetermined state where the main operator is not possible to perform an operation or is not performing an operation. The input unit is configured to accept an operational input based on a voice of the main operator, as well as to accept an operational input based on a voice of a passenger of the vehicle in a case where the determination unit has determined that the main operator is in the predetermined state.

Abstract: Dialog acts (e.g., questions) are selected for voice browsing by a machine learning model trained to identify a dialog act that is most likely to lead to a desired outcome. When an invocation is received from a user, a context of the invocation is determined, and a pool of dialog acts is scored based on the context by a machine learning model. Dialog acts are selected from the pool and presented to the user in accordance with a randomization policy. Data regarding the dialog acts and their success in achieving a desired outcome is used to train one or more machine learning models to select dialog acts in response to invocations.

Abstract: A sentiment analysis of a chat session in which a plurality of chat messages are posted is performed. Based on the sentiment analysis, at least one emotive model is derived for the chat session. A sentiment of users in the chat session can be determined using the emotive model. A user composing a new chat message for the chat session can be monitored. Based on the monitoring, an impact on the sentiment of the users in the chat session by the new chat message can be determined. A client device can be initiated to display the impact on the sentiment of the users in the chat session by the new chat message before the new chat message is posted in the chat session.

Abstract: An electronic device is provided. The electronic device includes a microphone, a display, a camera, a processor, and a memory. The processor is configured to receive a first utterance input through the microphone. The processor is also configured to obtain first recognized data from a first image displayed on the display or stored in the memory. The processor is further configured to store the first recognized data in association with the first utterance input when the obtained first recognized data matches the first utterance input. Additionally, the processor is configured to activate the camera when the first recognized data does not match the first utterance input. The processor is also configured to obtain second recognized data from a second image collected through the camera and store the second recognized data in association with the first utterance input when the obtained second recognized data matches the first utterance input.

Abstract: Systems and methods for neural network training utilizing loss functions reflecting neighbor token dependencies.

Abstract: Embodiments of the present disclosure are directed to a system, methods, and computer-readable media for facilitating stylistic expression transfers in machine translation of source sequence data. Using integrated loss functions for style transfer along with content preservation and/or cross entropy, source sequence data is processed by an autoencoder trained to reduce loss values across the loss functions at each time step encoded for the source sequence data. The target sequence data generated by the autoencoder therefore exhibits reduced loss values for the integrated loss functions at each time step, thereby improving content preservation and providing for stylistic expression transfer.

Abstract: Systems and methods for recognizing and executing spoken commands using speech recognition. Exemplary implementations may: store actionable phrases; obtain audio information representing sound captured by a mobile client computing platform associated with a user; detect any spoken instances of a predetermined keyword present in the sound represented by the audio information; perform speech recognition on the sound represented by the audio information; identify an utterance of an individual actionable phrase in speech temporally adjacent to the spoken instance of the predetermined keyword that is present in the sound represented by the audio information; perform natural language processing to identify an individual command uttered temporally adjacent to the spoken instance of the predetermined keyword that is present in the sound represented by the audio information; and effectuate performance of instructions corresponding to the command.

Abstract: There is provided an information processing device including an analysis unit configured to analyze a character string indicating contents of utterance obtained as a result of speech recognition, and a display control unit configured to display the character string indicating the contents of the utterance and an analysis result on a display screen.

Abstract: Embodiments of the invention may be used to implement a rate converter that includes: 6 channels in forward (audio) path, each channel having a 24-bit signal path per channel, an End-to-end SNR of 110 dB, all within the 20 Hz to 20 KHz bandwidth. Embodiment may also be used to implement a rate converter having: 2 channels in a reverse path, such as for voice signals, 16-bit signal path per channel, an End-to-end SNR of 93 dB, all within 20 Hz to 20 KHz bandwidth. The rate converter may include sample rates such as 8, 11.025, 12, 16, 22.05, 24, 32 44.1, 48, and 96 KHz. Further, rate converters according to embodiments may include a gated clock in low-power mode to conserve power.

Abstract: In accordance with an aspect of the disclosure, an electronic device comprises a first audio receiving circuit; a communication circuit; at least one processor operatively connected to the first audio receiving circuit and the communication circuit; and a memory operatively connected to the at least one processor, wherein the memory stores one or more instructions that, when executed, cause the at least one processor to: obtain first audio data, wherein the first audio data is based on a user utterance recorded by an external electronic device, through the communication circuit; convert the first audio data into second audio data, using a first transfer function of the first audio receiving circuit; and generate a wakeup model using the second audio data, the wakeup model configured to verify a wakeup word associated with the first audio data.

Abstract: Various arrangements for metadata tagging of video content are presented. A request to add a metadata tag to be linked with a video content instance may be received. A metadata integration database to link the spoken metadata tag with the video content instance may be updated.

Abstract: An artificial intelligence device may receive first voice data corresponding to first voice uttered by a user from a first peripheral device, acquire a first intention corresponding to the first voice data, transmit a first search result corresponding to the first intention to the first peripheral device, receive second voice data corresponding to second voice uttered by the user from a second peripheral device, acquire a second intention corresponding to the received second voice data, and transmit a search result corresponding to the second intention to the second peripheral device depending on whether the second intention is an interactive intention associated with the first intention.

Abstract: Techniques performed by a data processing system for training a Recurrent Neural Network Transducer (RNN-T) herein include encoder pretraining by training a neural network-based token classification model using first token-aligned training data representing a plurality of utterances, where each utterance is associated with a plurality of frames of audio data and tokens representing each utterance are aligned with frame boundaries of the plurality of audio frames; obtaining first cross-entropy (CE) criterion from the token classification model, wherein the CE criterion represent a divergence between expected outputs and reference outputs of the model; pretraining an encoder of an RNN-T based on the first CE criterion; and training the RNN-T with second training data after pretraining the encoder of the RNN-T. These techniques also include whole-network pre-training of the RNN-T.

Abstract: A system according to various embodiments may comprise: a first electronic device including a user interface and a communication circuit; at least one processor which is a part of the first electronic device or remotely communicates with the first electronic device; and at least one memory residing in the first electronic device or residing outside the first electronic device.

Abstract: Disclosed herein are methods of diarizing audio data using first-pass blind diarization and second-pass blind diarization that generate speaker statistical models, wherein the first pass-blind diarization is on a per-frame basis and the second pass-blind diarization is on a per-word basis, and methods of creating acoustic signatures for a common speaker based only on the statistical models of the speakers in each audio session.

Abstract: An example computing system that maintains a playlist of media items for playback by one or more playback devices receives, from a first playback device, a request to play back the playlist and first operational data and also receives, from a second playback device, a request to play back the playlist and second operational data. The computing system provides the media items in the playlist to the first and second playback devices for substantially synchronous playback and, based on the first and second operational data, determines respective first and second advertisements. The computing system provides the first advertisement to the first playback device for playback during a given time slot designated for advertisements, and also provides the second advertisement to the second playback device for playback during the given time slot, such that the second playback device plays the second advertisement while the first playback device plays the first advertisement.

Abstract: A method for generating human-machine hybrid predictions of answers to forecasting problems includes: parsing text of an individual forecasting problem to identify keywords; generating machine models based on the keywords; scraping data sources based on the keywords to collect scraped data relevant to the individual forecasting problem; providing the scraped data to the machine models; receiving machine predictions of answers to the individual forecasting problem from the machine models based on the scraped data; providing, by the computer system via a user interface, the scraped data to human participants; receiving, by the computer system via the user interface, human predictions of answers to the individual forecasting problem from the human participants; aggregating the machine predictions with the human predictions to generate aggregated predictions; and generating and outputting a hybrid prediction based on the aggregated predictions.

Abstract: A voice recognition system includes a server storing a plurality of manuals and a display apparatus transmitting, when a spoken voice of a user is recognized, characteristic information and a spoken voice signal corresponding to the spoken voice to the server, the characteristic information is characteristic information of the display apparatus, the server transmits a response signal to the spoken voice signal to the display apparatus based on a manual corresponding to the characteristic information among the plurality of manuals, and the display apparatus processes an operation corresponding to the received response signal; as a result, user convenience increases.

Abstract: A method, an electronic device and computer readable medium for dialogue breakdown detection are provided. The method includes obtaining a verbal input from an audio sensor. The method also includes generating a reply to the verbal input. The method additionally includes identifying a local context from the verbal input and a global context from the verbal input, additional verbal inputs previously received by the audio sensor, and previous replies generated in response to the additional verbal inputs. The method further includes identifying a dialogue breakdown in response to determining that the reply does not correspond to the local context and the global context. In addition, the method includes generating sound corresponding to the reply through a speaker when the dialogue breakdown is not identified.

Abstract: A method of controlling an intelligent security device can include capturing a video; collecting voice information included in the video; in response to determining that the voice information includes a wake-up word corresponding to a predetermined basic wake-up word for the intelligent security device, transmitting a spoken utterance included in the voice information to a smart device; receiving a command from the smart device, the command being generated based on information related to the spoken utterance; and executing an operation of the intelligent security device based on the command.