Abstract: A message may be suggested to a user participating in a conversation using one or more neural networks where the suggested message is adapted to the preferences or communication style of the user. The suggested message may be adapted to the user with a user embedding vector that represents the preferences or communication style of the user in a vector space. To suggest a message to the user, a conversation feature vector may be computed by processing the text the conversation with a neural network. A context score may be computed for one or more designated messages, where the context score is computed by processing the user embedding vector, the conversation feature vector, and a designated message feature vector with a neural network. A designated message may be selected as a suggested message for the user using the context scores. The suggestion may then presented to the user.

Abstract: A data structuring device 1 includes an acquisition unit 2 and a modeling unit 3. The acquisition unit 2 acquires branch expression data. Branch expression data is data that includes a branch expression, which is an utterance expression in a conversation for which multiple mutually different replies can be assumed. The conversation data is converted to a graph structure in which the content of the conversation is divided into multiple nodes, and the multiple nodes are connected by means of edges. On the basis of the branch expression data obtained by the acquisition unit 2, the modeling unit 3 carries out a process on the multiple items of conversation data in the graph structure in which nodes that share a common branch expression are combined in one node, thereby modeling the multiple items of conversation data in one tree structure.

Abstract: The present disclosure discloses an information processing method, a service providing system, and a computer storage medium. The method includes: assigning at least one first account for a multimedia device; assigning a plurality of second accounts associated with the first account for the multimedia device, each second account representing one user identifier; acquiring a voice instruction; determining a currently used second account according to the voice instruction; searching for user preference information associated with the currently used second account; and providing, based on the user preference information and service request information determined based on the voice instruction, a first service to the user identifier represented by the currently used second account.

Abstract: Systems and methods are described for generating learning roadmaps from unstructured information. The systems and methods may provide for extracting a plurality of named entities from one or more corpora of information, constructing a graph based on the named entities, inducing a subgraph from the graph based on a target named entity, wherein the subgraph includes a subset of the named entities, ordering the subset of the named entities based on the subgraph, and generating a learning roadmap for the target named entity based on the ordering.

Abstract: A method includes identifying, by an electronic device, one or more segments within a first audio recording that includes one or more non-speech segments and one or more speech segments. The method also includes generating, by the electronic device, one or more synthetic speech segments that include natural speech audio characteristics and that preserve one or more non-private features of the one or more speech segments. The method also includes generating, by the electronic device, an obfuscated audio recording by replacing the one or more speech segments with the one or more synthetic speech segments while maintaining the one or more non-speech segments, wherein the one or more synthetic speech segments prevent recognition of some content of the obfuscated audio recording.

Abstract: A system and method that enhances spoken utterances and provides entertainment by capturing one or more microphone signals containing echo and decomposing the one or more microphone signals into a plurality of signal paths through a synthesizer that adds or makes non-linear modifications to some of the captured one or more microphone signals. The system and method and estimates multiple echo paths from each of the one the one or more microphones. The system and method processes the captured microphone signals in response to the estimated plurality of echo paths by subtracting the echo contributions of each of the plurality of echo paths from the captured one or more microphone signals. The system and method also provide signal separation and post processing functions that renders speech recognition gaming applications.

Abstract: Techniques described herein are directed to analyzing intellectual-property data according to provide various intellectual property related services to organizations. In particular implementations, information related to products and/or services may be obtained from a number of data sources. Additionally, information related to intellectual-property assets, such as patents, trademarks, copyrights, trade secrets, and know-how, may be obtained. In various situations, the intellectual-property assets may be mapped to respective products and/or services. The mappings between the products and/or services and intellectual-property assets may be used to provide intellectual property related services that correspond to the intellectual-property assets, such as valuation services, strategy-related services, or risk-related services.

Abstract: In a general aspect, an apparatus for detecting speech can include a signal conditioning stage that receives a signal corresponding with acoustic energy, filters the received signal to produce a speech-band signal, calculates a first sequence of energy values for the received signal and calculates a second sequence of energy values for the speech-band signal. The apparatus can also include a detection stage including a plurality of speech and noise differentiators. The detection stage can being configured to receive the first and second sequences of energy values and, based on the first sequence of energy values and the second sequence of energy values, provide, for each speech and noise differentiator of the plurality of speech and noise differentiators, a respective speech-detection indication signal.

Abstract: Aspects of the disclosure relate to cognitive automation-based engine processing on audio files and streams received from meetings and/or telephone calls. A noise mask can be applied to enhance the audio. Real-time speech analytics separate speech for different speakers into time-stamped streams, which are transcribed and merged into a combined output. The output is parsed by analyzing the combined output for correct syntax, normalized by breaking the parsed data into record groups for efficient processing, validated to ensure that the data satisfies defined formats and input criteria, and enriched to correct for any errors and to augment the audio information. Notifications based on the enriched data may be provided to call or meeting participants.

Abstract: Systems and methods are described for generating learning roadmaps from unstructured information. The systems and methods may provide for extracting a plurality of named entities from one or more corpora of information, constructing a graph based on the named entities, inducing a subgraph from the graph based on a target named entity, wherein the subgraph includes a subset of the named entities, ordering the subset of the named entities based on the subgraph, and generating a learning roadmap for the target named entity based on the ordering.

Abstract: A system and method are presented for F0 transfer learning for improving F0 prediction with deep neural network models. Larger models are trained using long short-term memory (LSTM) and multi-layer perceptron (MLP) feed-forward hidden layer modeling. The fundamental frequency values for voiced and unvoiced segments are identified and extracted from the larger models. The values for voiced regions are transferred and applied to training a smaller model and the smaller model is applied in the text to speech system for real-time speech synthesis output.

Abstract: An apparatus and a method include receiving an input audio signal to be processed by a multi-band synchronized neural vocoder. The input audio signal is separated into a plurality of frequency bands. A plurality of audio signals corresponding to the plurality of frequency bands is obtained. Each of the audio signals is downsampled, and processed by the multi-band synchronized neural vocoder. An audio output signal is generated.

Abstract: A control method outputs audio indicating content of an operation of a transmission device connected to a reception device. The control method includes accepting the operation of the transmission device, generating operation data indicating the content of the operation by the transmission device, transmitting the operation data from the transmission device to the reception device, generating audio data indicating the content of the operation based on the operation data by the reception device, and outputting the audio indicated by the audio data by the reception device.

Abstract: Mechanisms are provided to implement a generative adversarial network (GAN) for natural language processing. With these mechanisms, a generator neural network of the GAN is configured to generate a bag-of-ngrams (BoN) output based on a noise vector input and a discriminator neural network of the GAN is configured to receive a BoN input, where the BoN input is either the BoN output from the generator neural network or a BoN input associated with an actual portion of natural language text. The mechanisms further configure the discriminator neural network of the GAN to output an indication of a probability as to whether the input BoN is from the actual portion of natural language text or is the BoN output of the generator neural network.

Abstract: Representative features representing actions and states described in text corpus comprising unstructured natural language data are generated. State properties associated with the states from the representative features and action properties associated with the actions from the representative features are generated. A state transition matrix with initial configurations is generated based on the state space and action space properties. A computer environment is bootstrapped based on the state space, the action space and the initial state transition matrix.

Abstract: A method, system and storage medium for solving echo cancellation failure; method comprises: starting echo cancellation algorithm for convergence of self-adapting filter to stabilize coefficient of filter function (S100); presetting maximum amplitude value for signal intensity that microphone can receive, and detecting if microphone signal lower than maximum amplitude value (S200); prompting user or television system automatically to turn down television volume when microphone received signal larger than or equal to maximum amplitude value, to ensure microphone received signal intensity lower than maximum amplitude value (S300); echo cancellation being carried out without distortion of signals, wherein coefficient of self-adapting filter function W(n) may be continually kept stable, avoiding self-adapting filter converging multiple times due to signal distortion (S400).

Abstract: A computer system that generates output speech is described. During operation, the computer system may receive an input associated with a type of interaction. Then, the computer system may generate, using a voice synthesis engine, the output speech corresponding to an individual based at least in part on the input, where the voice synthesis engine predicts positions and duration of a prosodic characteristic of speech by the individual, and selectively adds the prosodic characteristic of the speech by the individual in the output speech based at least in part on the prediction. Note that the prosodic characteristic may include: pauses in the speech by the individual, and/or disfluences in the speech by the individual.

Abstract: An artificial intelligence apparatus for recognizing speech by correcting misrecognized word includes a microphone and a processor. The processor is configured to obtain, via the microphone, speech data including speech of a user, convert the speech data into text by using an acoustic model and a language model, determine whether an uncertain recognition exists in an acoustic recognition result according to the acoustic model, determine whether the converted text is a normal sentence by using a natural language processing model if an uncertain recognition exists in the acoustic recognition result, determine a sentence most similar to the converted text among sentences pre-learned by using the language model if the converted text is not a normal sentence, replace the converted text with the determined most similar sentence, and generate a speech recognition result corresponding to the speech data by using the converted text.

Abstract: Techniques facilitating detection of conversation threads in unstructured channels are provided. A system can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise an extraction component that employs a model to detect conversation messages based on a defined confidence level and assigns the conversation messages to respective conversation thread categories. The computer executable components also can comprise a model component that trains the model on conversation messages that comprise respective text data, wherein the model is trained to detect the respective text data to the defined confidence level.

Abstract: The disclosure relates to the technical field of lifting mechanisms, and in particular to a voice control system and a method for moving columns, comprising a lifting mechanism, a controller, a storage module, and a voice collection module. The lifting mechanism, the storage module and the voice collection module are all electrically connected to the controller. A control word and at least two wake-up words are stored in the storage module. The system further includes a selection module for selecting the wake-up word and a display module for displaying the current wake-up word. Both the selection module and the display module are electrically connected to the controller. The likelihood of unwanted triggering can be decreased by the voice control system and method.