Abstract: A method for employing a temporal context-aware question routing model (TCQR) in multiple granularities of temporal dynamics in community-based question answering (CQA) systems is presented. The method includes encoding answerers into temporal context-aware representations based on semantic and temporal information of questions, measuring answerers expertise in one or more of the questions as a coherence between the temporal context-aware representations of the answerers and encodings of the questions, modeling the temporal dynamics of answering behaviors of the answerers in different levels of time granularities by using multi-shift and multi-resolution extensions, and outputting answers of select answerers to a visualization device.

Abstract: Based on a candidate set of translations produced by a neural network based machine learning model, a mapping data structure such as a statistical phrase table is generated. The mapping data structure is analyzed to obtain a quality metric of the neural network based model. One or more operations are initiated based on the quality metric.

Abstract: Techniques are described herein for training and evaluating machine learning (ML) models for document processing computing applications using generalized vocabulary tokens. In some embodiments, an ML system determines a set of tokens for non-textual content in a plurality of documents. The ML system generates a fixed-length vocabulary that includes the set of tokens for the non-textual content. The ML system further generates for each respective document in a training dataset of documents, a respective feature vector based at least in part on which tokens in the fixed-length vocabulary occur in the respective document. The ML system trains a ML model based at least in part on the respective feature vector for each respective document in the training dataset.

Abstract: Decorrelating an input signal includes allpass filtering to phase shift the first input signal by a phase shift, the allpass filtering comprising filtering with one or more subsequent controllable allpass filter stages, each controllable allpass filter stage having a filter quality and a cut-off frequency. Decorrelating further includes controlling at least one of the filter quality and the cut-off frequency of the controllable allpass filter stages to change over time.

Abstract: A computing machine receives an input comprising unstructured text. The computing machine identifies, within the unstructured text, one or more entities using a named entity recognition (NER) engine in a trained machine learning model. The trained machine learning model embeds tokens from the text into a vector space and uses generated embeddings to identify one or more tokens as being associated with the one or more entities. The computing machine determines, using the trained machine learning model that identifies the one or more entities and based on the embedded tokens, an assertion applied, within the text, to at least one entity. The assertion is represented as a vector in a multi-dimensional space. Each dimension corresponds to a part of the assertion. The trained machine learning model is a span-level model that both identifies the one or more entities and determines the assertion based on candidate spans of tokens.

Abstract: A system and method for assisting with interactions between agents and customers using an artificially intelligent assistant is disclosed. The artificially intelligent assistant monitors interactions between agents and customers and identifies assistive actions to be taken that increase efficiency of the interaction as well as customer satisfaction. The artificially intelligent agent can also identify new communication modes appropriate for assistive actions, allowing agents to seamlessly communicate with customers over a wide range of different communication modes, such as phone calls, texts, emails and other messaging applications.

Abstract: A method, computer program product, and/or computer system generate a first adversarial statement via: (1) receiving a question and an original context for the question; (2) converting the question into a statement with a placeholder answer; (3) picking randomly an answer entity from a training text corpus; (4) replacing the placeholder answer with the randomly picked answer entity; and (5) leaving a correct question entity in the statement. The first adversarial statement is inserted into the original context to form a first adversarial context. The question and the first adversarial context as a first pair and the question and the original context as a second pair are input into a question-answer dialog system to train the question-answer dialog system.

Abstract: A virtual assistant system for communicating with customers uses human intelligence to correct any errors in the system AI, while collecting data for machine learning and future improvements for more automation. The system may use a modular design, with separate components for carrying out different system functions and sub-functions, and with frameworks for selecting the component best able to respond to a given customer conversation. The system may have agent assistance functionality that uses natural language processing to identity concepts in a user conversation and to illustrate that concepts within a graphical user interface of a human agent so that the human agent can more accurately and more rapidly assist the user in accomplishing the user's conversational objectives.

Abstract: A system and method for dynamically providing safe call back numbers to use to respond to an inbound communication, the method comprising parsing message records from a message server, analyzing the message records for untrustworthy phone numbers by comparing content of the message records to a reference data set that is retrieved from a database, the reference data set including genuine and fraud data wherein the genuine and fraud data includes entities and contact information corresponding to the entities, determining untrustworthy phone numbers in the message records from the analysis, and generating remedy actions based on the determination of the untrustworthy phone numbers.

Abstract: A communication session may be established between a customer device and a customer service representative device. An artificial intelligence assistant may access the communication session and receive an input. The artificial intelligence assistant may process in the input. The artificial intelligence assistant may determine one or more characteristics of the input. The artificial intelligence assistant may determine an output based on the characteristics of the input. The artificial intelligence assistant may communicate the determined output to the customer service representative.

Abstract: A method includes generating a synthesized non-reference high-band channel based on a non-reference high-band excitation corresponding to a non-reference target channel. The method further includes estimating one or more spectral mapping parameters based on the synthesized non-reference high-band channel and a high-band portion of the non-reference target channel. The method also includes applying the one or more spectral mapping parameters to the synthesized non-reference high-band channel to generate a spectrally shaped synthesized non-reference high-band channel. The method further includes generating an encoded bitstream based on the one or more spectral mapping parameters and the spectrally shaped synthesized non-reference high-band channel.

Abstract: At least some embodiments, a system includes a memory, and a processor configured to convert an audio stream of a speech of a customer during a customer call session into customer-originated text. The customer-originated text is displayed in a first chat interface. A request from a first call center agent is sent to a second call center agent via the first chat interface to interact with the customer during the customer call session and displayed in a second chat interface. The second agent is allowed to participate in the customer call session when the second call center agent accepts the request from the first call center agent. First agent-originated text and second agent-originated text during the customer call session is merged to form a combined agent-originated text and synthesized to computer-generated agent speech having a voice of a computer-generated agent based on the combined agent-originated text communicated to the customer over the voice channel.

Abstract: A hearing instrument includes: a first portion shaped and sized for placement at a pinna of a user's ear; and a second portion having an earpiece for placement in the user's ear canal; wherein the second portion also comprises a connector assembly configured for electrically coupling to the first portion, the connector assembly having a plurality of connector wires, the plurality of connector wires comprising a first connector wire; wherein the second portion also comprises a receiver or miniature loudspeaker for receipt of an audio drive signal through at least the first connector wire; and wherein the second portion also comprises a non-volatile memory circuit having a data interface configured for receipt and transmittal of module data, the non-volatile memory circuit configured to store the module data, wherein the stored module data at least comprises electroacoustic calibration parameter(s) of the receiver or the miniature loudspeaker.

Abstract: A number management unit (11) of a number management system (1) generates a transaction using number portability information including a hashed telephone number and connection destination information, newly generates and approves a block using multiple transactions, and stores the block in a number database (13) as a blockchain. A number resolution unit (12) hashes a telephone number that is a query target using a hash function, searches transactions in a blockchain using the hashed telephone number as a key, and sends connection destination information corresponding to the telephone number that is the query target as a response.

Abstract: Embodiments described herein provide a composed variational natural language generation (CLANG) model that is configured to generate training samples for few-shot intents. Specifically, the CLANG model may build connections between existing training samples of many-shot intents and new training samples of few-shot intents by modeling an intent as a combination of a domain and an action. In this way, the CLANG model transfers knowledge from existing many-shot intents to few-shot intents in natural language generation by learning how to compose utterances with many-shot intents and transferring such knowledge to few-shot intents.

Abstract: A number management unit (11) of a number management system (1) generates a transaction using number portability information including a telephone number and encrypted connection destination information, newly generates a block using multiple transactions, approves the generated block, and stores the block in a number database (13) as a blockchain. A number resolution unit (12) searches transactions in the blockchain using a telephone number that is a query target as a key, extracts and decrypts the encrypted connection destination information corresponding to the query telephone number, and sends the connection destination information corresponding to the query telephone number as a response.

Abstract: The disclosure provides a text recognition method, an electronic device, and a storage medium. The method includes: obtaining N segments of a sample text; inputting each of the N segments into a preset initial language model in sequence, to obtain first text vector information corresponding to the N segments; inputting each of the N segments into the initial language model in sequence again, to obtain second text vector information corresponding to a currently input segment; in response to determining that the currently input segment has the mask, predicting the mask according to the second text vector information and the first text vector information to obtain a predicted word at a target position corresponding to the mask; training the initial language model according to an original word and the predicted word to generate a long text language model; and recognizing an input text through the long text language model.

Abstract: Various examples are directed to systems and methods for detecting potentially fraudulent voice calls to a financial services institution. A computing system may receive an indication of a voice call placed by a voice caller to an operator. The computing system may generate a network address indicator describing a network location. The network address indicator may be provided to the voice caller. The computing system may receive an indication of a financial services account indicated by the voice caller. The computing system may also receive an indication of an access to the network location by a remote device. The computing system may determine, using the indication of the access to the network location, a first location associated with the remote device and determine that the first location does not match a second location associated with the financial services account. The computing system may generate an alert indicating that the voice call is potentially fraudulent.

Abstract: A method and apparatus for recognizing a user intention, a device, and a readable storage medium are disclosed. A specific implementation of the method includes: acquiring a plurality of rounds of dialogue data and a satisfaction degree of each round of dialogue data, selecting target dialogue data having a satisfaction degree meeting a set satisfaction condition; labeling input data in the target dialogue data with an intention label; and training the intention recognition model based on the input data in the target dialogue data and the intention label of the input data, so that the trained intention recognition model performs intention recognition on new input data.

Abstract: Methods, devices, and systems for contactless cough detection and attribution are presented herein. Audio data may be received using a microphone. A cough may be identified as having occurred based on the received audio data. Radar data may be received indicative of reflected radio waves from a radar sensor. A state analysis process may be performed using the received radar data. The detected cough may be attributed to a particular user based at least in part on the state analysis process performed using the radar data.