Abstract: The present invention relates to method and device for text-enhanced knowledge graph joint representation learning, the method at least comprises: learning a structure vector representation based on entity objects and their relation linking in a knowledge graph and forming structure representation vectors; discriminating credibility of reliable feature information and building an attention mechanism model, aggregating vectors of different sentences and obtain association-discriminated text representation vectors; and building a joint representation learning model, and using a dynamic parameter-generating strategy to perform joint learning for the text representation vectors and the structure representation vectors based on the joint representation learning model. The present invention selective enhances entity/relation vectors based on significance of associated texts, so as to provide improved semantic expressiveness, and uses 2D convolution operations to train joint representation vectors.

Abstract: A method of sampling output audio samples includes, during a packet loss concealment event, obtaining a sequence of previous output audio samples. At each time step during the event, the method includes generating a probability distribution over possible output audio samples for the time step. Each sample includes a respective probability indicating a likelihood that the corresponding sample represents a portion of an utterance at the time step. The method also includes determining a temperature sampling value based on a function of a number of time steps that precedes the time step, and an initial, a minimum, and a maximum temperature sampling value. The method also includes applying the temperature sampling value to the probability distribution to adjust a probability of selecting possible samples and randomly selecting one of the possible samples based on the adjusted probability. The method also includes generating synthesized speech using the randomly selected sample.

Abstract: In accordance with embodiments of the present disclosure, a method for processing audio information in an audio device may include reproducing audio information by generating an audio output signal for communication to at least one transducer of the audio device, receiving at least one input signal indicative of ambient sound external to the audio device, detecting from the at least one input signal a near-field sound in the ambient sound, and modifying a characteristic of the audio information reproduced to the at least one transducer in response to detection of the near-field sound.

Abstract: Determining slot value(s) based on received natural language input and based on descriptor(s) for the slot(s). In some implementations, natural language input is received as part of human-to-automated assistant dialog. A natural language input embedding is generated based on token(s) of the natural language input. Further, descriptor embedding(s) are generated (or received), where each of the descriptor embeddings is generated based on descriptor(s) for a corresponding slot that is assigned to a domain indicated by the dialog. The natural language input embedding and the descriptor embedding(s) are applied to layer(s) of a neural network model to determine, for each of the slot(s), which token(s) of the natural language input correspond to the slot. A command is generated that includes slot value(s) for slot(s), where the slot value(s) for one or more of slot(s) are determined based on the token(s) determined to correspond to the slot(s).

Abstract: Provided are a voice synthesis method, an apparatus, a device, and a storage medium, involving obtaining text information and determining characters in the text information and a text content of each of the characters; performing a character recognition on the text content of each of the characters, to determine character attribute information of each of the characters; obtaining speakers in one-to-one correspondence with the characters according to the character attribute information of each of the characters, where the speakers are pre-stored pronunciation object having the character attribute information; and generating multi-character synthesized voices according to the text information and the speakers corresponding to the characters of the text information. These improve pronunciation diversities of different characters in the synthesized voices, improve an audience's discrimination between different characters in the synthesized voices, and thereby improve experience of a user.

Abstract: In an example, a text sentence comprising a plurality of words is obtained. Each of the plurality of words is passed through a deep compositional character-to-word model to encode character-level information of each of the plurality of words into a character-to-word expression. The character-to-word expressions are combined with pre-trained word embeddings. The combined character-to-word expressions and pre-trained word embeddings are fed into one or more bidirectional long short-term memories to learn contextual information for each of the plurality of words. Then, sequential conditional random fields are applied to the contextual information for each of the plurality of words.

Abstract: Techniques regarding speech characterization are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a speech analysis component that can determine a condition of an origin of an audio signal based on a difference between a first feature of the audio signal and a second feature of a synthesized reference audio signal.

Abstract: In one example, a system can execute a first machine-learning model to determine an overall classification for a textual dataset. The system can also determine classification scores indicating the level of influence that each token in the textual dataset had on the overall classification. The system can select a first subset of the tokens based on their classification scores. The system can also execute a second machine-learning model to determine probabilities that the textual dataset falls into various categories. The system can determine category scores indicating the level of influence that each token had on a most-likely category determination. The system can select a second subset of the tokens based on their category scores. The system can then generate a first visualization depicting the first subset of tokens color-coded to indicate their classification scores and a second visualization depicting the second subset of tokens color-coded to indicate their category scores.

Abstract: Concealing a lost audio frame of a received audio signal is provided by performing a sinusoidal analysis (81) of a part of a previously received or reconstructed audio signal, wherein the sinusoidal analysis involves identifying frequencies of sinusoidal components of the audio signal, applying a sinusoidal model on a segment of the previously received or reconstructed audio signal, wherein said segment is used as a prototype frame in order to create a substitution frame for a lost audio frame, and creating the substitution frame (83) for the lost audio frame by time-evolving sinusoidal components of the prototype frame, up to the time instance of the lost audio frame, in response to the corresponding identified frequencies.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing speech separation. One of the methods includes obtaining a recording comprising speech from a plurality of speakers; processing the recording using a speaker neural network having speaker parameter values and configured to process the recording in accordance with the speaker parameter values to generate a plurality of per-recording speaker representations, each speaker representation representing features of a respective identified speaker in the recording; and processing the per-recording speaker representations and the recording using a separation neural network having separation parameter values and configured to process the recording and the speaker representations in accordance with the separation parameter values to generate, for each speaker representation, a respective predicted isolated audio signal that corresponds to speech of one of the speakers in the recording.

Abstract: An electronic device and method of recognizing an audio scene are provided.

Abstract: The disclosure describes one or more embodiments of an acoustic improvement system that accurately and efficiently determines and provides actionable acoustic improvement suggestions to users for digital audio recordings via an interactive graphical user interface. For example, the acoustic improvement system can assist users in creating high-quality digital audio recordings by providing a combination of acoustic quality metrics and actionable acoustic improvement suggestions within the interactive graphical user interface customized to each digital audio recording. In this manner, all users can easily and intuitively utilize the acoustic improvement system to improve the quality of digital audio recordings.

Abstract: Disclosed are a method and an apparatus for editing audio, an electronic device and a storage medium. The method includes: acquiring a modified text obtained by modifying a known original text of an audio to be edited according to a known text for modification; predicting a duration of an audio corresponding to the text for modification; adjusting a region to be edited of the audio to be edited according to the duration of the audio corresponding to the text for modification, to obtain an adjusted audio to be edited; obtaining, based on a pre-trained audio editing model, an edited audio according to the adjusted audio to be edited and the modified text. In the present disclosure, the edited audio obtained by the audio editing model sounds natural in the context, and supports the function of synthesizing new words that do not appear in the corpus.

Abstract: The technology of the present application provides software as a service (SaaS) executing on a server in a cloud or network. The SaaS receives data from a mobile device of a user over the network. The SaaS processes the data and returns the processed data to a client application executing on a client device of the user, which user is the same as the user of the mobile device wherein there is no direct communication link, wireless or wired, between the mobile device and the client device. In one aspect, the technology of the present application provides the mobile device as a smartphone and a microphone application to be executed on the smartphone.

Abstract: A method, computer program product, and computing system for obtaining calibration information for a three-dimensional space incorporating an ACI system; and processing the calibration information to calibrate the ACI system.

Abstract: A speech-processing system capable of receiving and processing audio data to determine if the audio data includes speech that was intended for the system. Non-system directed speech may be filtered out while system-directed speech may be selected for further processing. A system-directed speech detector may use a trained machine learning model (such as a deep neural network or the like) to process a feature vector representing a variety of characteristics of the incoming audio data, including the results of automatic speech recognition and/or other data. Using the feature vector the model may output an indicator as to whether the speech is system-directed. The system may also incorporate other filters such as voice activity detection prior to speech recognition, or the like.

Abstract: A method and computing platform to imitate human conversational response as a context transitions across multiple channels (e.g., chat, messaging, email, voice, third party communication, etc.) where inputs to the system are categorized into identified speech acts and physical acts, and a conversational bot is associated to the channels. In this approach, a data model associated with a multi-turn conversation is provided. The data model comprises an observation history, wherein an observation in the observation history includes an identification of a channel in which the observation originates. As turns are added to the multi-turn conversation, a conversational context across multiple channels is persisted using the data model. Using this approach, an AI-supported conversation started in one channel can move to another conversation channel while maintaining the context of the conversation intact and coherent.

Abstract: A method and device for training an acoustic model are provided. The method comprises determining a plurality of tasks for training an acoustic model, obtaining resource occupancies of nodes participating in the training of the acoustic model, and distributing the tasks to the nodes according to the resource occupancies of the nodes and complexities of the tasks. By using computational resources distributed at multiple nodes, tasks for training an acoustic model are performed in parallel in a distributed manner, so as to improve training efficiency.

Abstract: Computer implemented systems and methods of processing clinical documentation for a multi-axial coding scheme include inputting clinical documentation from memory operatively coupled with a computer system, and executing a natural language processor configured to process narrative text in the clinical documentation. The processor segments the narrative text based on boundaries defined in the clinical documentation, sequences words in the narrative text based on the segmentation, and maps the sequenced words to semantic objects in an ontology database. The ontology defines classes of semantic objects and relationships between them, corresponding to the multi-axial coding scheme. The semantic objects are converted into characters and output into slots in a medical code, with the characters positioned in the slots based on the multi-axial coding scheme.

Abstract: A method and method for natural language generation employ a natural language generation model which has been trained to assign an utterance label to a new text sequence, based on features extracted from the text sequence, such as parts-of-speech. The model assigns an utterance label to the new text sequence, based on the extracted features. The utterance label is used to guide the generation of a natural language utterance, such as a question, from the new text sequence. The system and method find application in dialog systems for generating utterances, to be sent to a user, from brief descriptions of problems or solutions in a knowledge base.