Abstract: Methods and systems for attention behavioral analysis for a conversational question and answer system are disclosed. A multi-modality input is selected from a plurality of multimodality conversations among two or more users. The system annotates the first modality inputs and at least one attention region in the first modality input corresponding to a set of entities and semantic relationships in a unified modality is identified by a discrete aspect of information bounded by the attention elements. The system models the representations of the multimodality inputs at different levels of granularity, which includes entity level, turn level, conversational level. The method proposed uses a network that consists of multilevel encoder-decoder architecture that is used to determine unified focalized attention, analyze and construct one or more responses for one or more turns in a conversation.

Abstract: A multi-tenant system performs custom configuration of a tenant-specific chatbot to process and act upon natural language requests. The multi-tenant system configures the tenant-specific chatbots without requiring tenant-specific training. The multi-tenant system providing a user interface for configuring a tenant-specific set of permitted actions. The multi-tenant system determines a set of example phrases for each of the selected permitted actions. The multi-tenant system receives a natural language request from a user and identifies the action that the user wants to perform. The multi-tenant system uses a neural network to compare the natural language request with example phrases to identify an example phrase that matches the natural language request. The multi-tenant system performs the action corresponding to the matching example phrase.

Abstract: Disclosed herein is an artificial intelligence apparatus including an input interface configured to receive speech data, and a processor configured to detect a non-utterance interval included in the speech data and determine presence/absence of a second utterance after the non-utterance interval according to characteristics of a first utterance before the non-utterance interval, when the non-utterance interval exceeds a set time.

Abstract: A computer-implemented method for speech synthesis, a computer device, and a non-transitory computer readable storage medium are provided. The method includes: obtaining a speech text to be synthesized; obtaining a Mel spectrum corresponding to the speech text to be synthesized according to the speech text to be synthesized; inputting the Mel spectrum into a complex neural network, and obtaining a complex spectrum corresponding to the speech text to be synthesized, wherein the complex spectrum comprises real component information and imaginary component information; and obtaining a synthetic speech corresponding to the speech text to be synthesized, according to the complex spectrum. The method can efficiently and simply complete speech synthesis.

Abstract: This disclosure describes techniques that include identifying sensitive information from any appropriate set of data, such as data produced by operations of a business or organization. In one example, this disclosure describes a method that includes receiving text data containing sensitive information, including structured sensitive information and unstructured sensitive information; applying a rule-based model to identify the structured sensitive information in the text data; applying a machine learning model to identify the unstructured sensitive information in the text data, wherein the machine learning model has been trained to identify unstructured sensitive information in text; and generating output text data from the text data by modifying the structured sensitive information identified by the rule-based model and the unstructured sensitive information identified by the machine learning model.

Abstract: Among other things, a developer of an interaction application for an enterprise can create items of content to be provided to an assistant platform for use in responses to requests of end-users. The developer can deploy the interaction application using defined items of content and an available general interaction model including intents and sample utterances having slots. The developer can deploy the interaction application without requiring the developer to formulate any of the intents, sample utterances, or slots of the general interaction model.

Abstract: Exemplary embodiments of restroom monitoring systems having a virtual assistants includes a communications gateway located in a restroom. The communications gateway having a processor, memory, short range communications circuitry, long range communications circuitry, a microphone and a speaker. The communications gateway containing logic for listening for a wake up word and upon detecting a wake up word, capturing a request, logic for processing the request to determine what request is being requested, logic for verifying the request with the requester, and one of a plurality of wave files and a voice synthesizer. The system further includes one or more dispensers located in the restroom. The one or more dispensers having short range communications circuitry for communicating status or product level to the communications gateway.

Abstract: An automatic agent may be improved through feedback. A user input may be received through a user interface. A plurality of current utterance variables may be obtained by tokenizing the user input. The automatic agent may execute a machine learning policy to generate a reply to the user input based on the plurality of current utterance variables. A different reply may be obtained in response to an indication that the reply will lead to a breakdown, wherein the breakdown comprises an unhuman response from the automatic agent according to the machine learning policy. The machine learning policy may be adjusted based on the plurality of current utterance variables and the different reply.

Abstract: A log processing device and a log processing method thereof are provided. The log processing device divides the original log data into a plurality of block data, transforms a numeric variable of each of the block data into a representative code, and determines whether to perform a combination process for continuous block data to generate a plurality of combinational block data according to a data integrity of each of the block data. The log processing data takes the combinational block data as a log template, and each of the combinational block data corresponds to an event.

Abstract: A sound association system identifies one or more aurally active words in digital text. Aurally active words refer to words that denote particular sounds. Context-based sounds corresponding to the one or more aurally active words are also identified. Each context-based sound is anchored to or associated with the corresponding one or more aurally active words and is played back when the digital text is played back or read, providing context-based background sounds associated with the one or more aurally active words. For example, a context-based sound can be played back at a higher volume when the one or more aurally active words are played back or read, and at a lower volume when other words of the digital text are played back or read.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for managing dialogs. In one aspect, a method includes receiving a request associated with a task from a user device; submitting the request to each of a plurality of distinct data providers; receiving a plurality of suggested dialog responses from two or more of the data providers; scoring the one or more suggested dialog responses based on one or more scoring factors; determining a particular dialog response to provide to the user based on the scoring; and providing the determined dialog response to the user device.

Abstract: A language determination model may be applied to select a first machine learning model or a second machine learning model to analyze the input text. The first machine learning model trained to analyze text in a first language, the second machine learning model trained to analyze text in a second language, and the input text may be in a third language. The language determination model may select the first machine learning model based on the first machine learning model having a better performance analyzing text in the third language than the second machine learning model. The language determination model may be updated based on an actual performance of the first machine learning model analyzing the input text. Moreover, the first machine learning model may be subject to additional training if the actual performance of the first machine learning model analyzing the input text is below a threshold value.

Abstract: A computing device receives a first audio waveform representing a first utterance and a second utterance. The computing device receives identity data indicating that the first utterance corresponds to a first speaker and the second utterance corresponds to a second speaker. The computing device determines, based on the first utterance, the second utterance, and the identity data, a diarization model configured to distinguish between utterances by the first speaker and utterances by the second speaker. The computing device receives, exclusively of receiving further identity data indicating a source speaker of a third utterance, a second audio waveform representing the third utterance. The computing device determines, by way of the diarization model and independently of the further identity data of the first type, the source speaker of the third utterance. The computing device updates the diarization model based on the third utterance and the determined source speaker.

Abstract: Methods for providing enhanced services to users participating in communication sessions (CS), via a virtual assistant, are disclosed. One method receives content that is exchanged by users participating in the CS. The content includes natural language expressions that encode a conversation carried out by users. The method determines content features based on natural language models. The content features indicate intended semantics of the natural language expressions. The method determines a relevance of the content and identifies portions of the content that are likely relevant to the user. Determining the relevance is based on the content features, a context of the CS, a user-interest model, and a content-relevance model of the natural language models. Identifying the likely relevant content is based on the determined relevance of the content and a relevance threshold. A summary of the CS is automatically generated from summarized versions of the likely relevant portions of the content.

Abstract: A method of speaker identification comprises receiving an audio signal representing speech; performing a first voice biometric process on the audio signal to attempt to identify whether the speech is the speech of an enrolled speaker; and, if the first voice biometric process makes an initial determination that the speech is the speech of an enrolled user, performing a second voice biometric process on the audio signal to attempt to identify whether the speech is the speech of the enrolled speaker. The second voice biometric process is selected to be more discriminative than the first voice biometric process.

Abstract: A high-band encoding/decoding method and device for bandwidth extension are provided. A high-band encoding method comprising the steps of: generating sub band-specific bit allocation information on the basis of a low-band envelope; determining, on the basis of the sub band-specific bit allocation information, the sub band requiring an envelope update in a high band; and generating, for the determined sub band, refinement data relating to the envelope update. A high-band decoding method comprising the steps of: generating sub band-specific bit allocation information on the basis of a low-band envelope; determining, on the basis of the sub band-specific bit allocation information, the sub band requiring an envelope update in a high band; and decoding, for the determined sub band, refinement data relating to the envelope update, thereby updating the envelope.

Abstract: The present teaching relates to automatic formality classification and transformation of online text items. In one example, a request is received for determining a formality level of a text item in an online communication. One or more linguistic features are extracted from the text item. Contextual information with respect to the online communication is extracted. A formality level of the text item is determined based on the one or more linguistic features and the contextual information. The formality level represents a degree of formality of the text item. The formality level is provided as a response to the request.

Abstract: A method and apparatus include receiving a text input that includes a sequence of text components. Respective temporal durations of the text components are determined using a duration model. A spectrogram frame is generated based on the duration model. An audio waveform is generated based on the spectrogram frame. Video information is generated based on the audio waveform. The audio waveform is provided as an output along with a corresponding video.

Abstract: A method for expanding an initial ontology via processing of communication data, wherein the initial ontology is a structural representation of language elements comprising a set of entities, a set of terms, a set of term-entity associations, a set of entity-association rules, a set of abstract relations, and a set of relation instances. A method for extracting a set of significant phrases and a set of significant phrase co-occurrences from an input set of documents further includes utilizing the terms to identify relations within the training set of communication data, wherein a relation is a pair of terms that appear in proximity to one another.

Abstract: Systems and methods for unified attribute extraction are disclosed. A set of product data including at least one text-based description of each of the products in the set of products is received and at least one closed list attribute is extracted from the at least one text-based description of each of the products. The at least one closed list attribute is extracted by an attribute extraction model configured to implement an extended conditional random field (XCRF) process. A set of attributes including each of the at least one closed list attributes extracted from the at least one text-based description of each of the products in the set of products is output.