Abstract: Matching content to third-party digital assistant actions using a bit vector is provided. A system receives an application with voice-assistant compatible actions. The system identifies the actions in the application. The system identifies content items provided by third-party computing devices. The system generates, via a machine learning model and performance data for the content items, bit vectors or the actions. The system selects, responsive to a request for content from a client device that executes an action of the application, a content item based on the bit vector that corresponds to the action.

Abstract: An information-seeking dialogue system can be trained using a pipeline process having stages, or components, of passage retrieval (selecting passages relevant to a query from a corpus or knowledge base), re-ranking, and generating a response to the query based on one or more of the re-ranked passages. Each stage, or component, of the pipeline can be individually optimized based on ground truth data.

Abstract: A system for preventing a voice activated response by a digital personal is provided. The system includes a digital personal assistant that has a memory storing program instructions. The system also includes one or more processors that, when executing the program instructions, are configured to identify an audible instruction based on a determined sound, and receive a block signal accompanying the determined sound. The one or more processors are also configured to prevent a response to the determined sound based on the block signal.

Abstract: An electronic device is provided and includes a processor and a memory operatively connected to the processor. The memory stores instructions that cause, when executed, the processor to acquire a first assistant result including data indicative of a first intent understood from an utterance of a first user, data indicative of an attribute of the utterance, first information provided to a terminal of the first user as a response of an artificial intelligence (AI) agent to the utterance, and a first parameter indicative of an attribute of the first information, to recognize the utterance as an information request utterance, based on the first intent and the attribute, and to track second information to be provided to the first user terminal as update information for the first information, based on the utterance being recognized as the information request utterance and the first parameter being identified as a specified type.

Abstract: Methods, systems, and non-transitory computer readable storage media are disclosed for correcting entity detection errors with entity correction and resolution in optical character recognition for digitization of physical documents. Specifically, the disclosed system utilizes named entity recognition to extract entities from character strings (e.g., words) in a digital text document. The disclosed system also tokenizes the character strings in the digital text document based on attributes of the character strings. Furthermore, the disclosed system compares the extracted entities and tokenized character strings to determine similarity metrics between the extracted entities and tokenized character strings. The disclosed system also compares extracted entities to character strings including special/numerical characters to determine similarity metrics indicating correlation probabilities between entities and character strings.

Abstract: Devices and methods for invoking digital assistants are provided. An example method includes displaying, on a display, a virtual reality (“VR”) environment, wherein the VR environment includes a first avatar representing a domain-specific digital assistant, and wherein the domain-specific digital assistant corresponds to a first domain; and while displaying the first avatar within the VR environment: receiving a user voice input; determining whether the user voice input represents an intent to invoke the domain-specific digital assistant or an intent to invoke a second digital assistant; and in accordance with a determination that the user voice input represents an intent to invoke the second digital assistant: invoking the second digital assistant; determining, based on the user voice input, a first digital assistant response using the second digital assistant; and outputting, via a speaker, the first digital assistant response.

Abstract: Disclosed is a computer program executable by one or more processors and stored in a computer-readable storage medium, the computer program causing the one or more processors to perform one or more operations below, the operations including: an operation of receiving first voice data from a first user terminal and receiving second voice data from a second user terminal; an operation of generating integrated voice data based on the first voice data and the second voice data; and an operation of generating the integrated voice data as a conversation record by using a voice recognition model.

Abstract: In a method for quantitatively evaluating a pronunciation of a speaker, an acoustic signal is received from the speaker, which represents an utterance spoken in a language by the speaker. The acoustic signal is segmented into segments at the phone, word or phrase level. The acoustic signal segments are each transcribed by a high fidelity transcriber into high fidelity transcription segments that each comprises a sequence of phones that represents how the native or normative speakers of the language would likely perceive a playback of the acoustic signal segment. Each of the high fidelity transcription segments are compared to a baseline to generate one or more pronunciation scores that evaluate the pronunciation of the speaker. The baseline may be generated by transcribing the acoustic signal into an adapted transcription that comprises a sequence of phones that represents how native or normative speakers of the language would likely speak the utterance.

Abstract: The present disclosure relates to method and apparatus for waking up a wearable device. The method includes: determining, based on a detected voice signal, whether the voice signal is from a wearer of the wearable device; determining, based on the voice signal, whether the voice signal belongs to a conversation between the wearer and the wearable device with a speech classification model; and waking up the wearable device when the voice signal is from the wearer and the voice signal belongs to the conversation between the wearer and the wearable device. The method and apparatus can wake up the wearable device during a normal interaction process between the wearer and the wearable device by using the method and apparatus, thereby realizing a natural wake-up process and improving the user experience of the wearable device.

Abstract: A method comprising detecting an activation of an intelligent assistant on an electronic device, waking up the intelligent assistant from a sleep mode in response to the activation, and determining an amount of vocabulary the intelligent assistant acts upon during a listening mode based on a type of the activation.

Abstract: A voice control system and a voice control method for an automatic door are provided. The voice control system includes a sound detection device, a storage device, a first determination circuit, a second determination circuit and a control circuit. The sound detection device detects a sound signal of a sound source, the storage device includes a voiceprint database that includes reference voiceprint features. The first determination circuit analyzes a voiceprint feature of the sound signal and compares the voiceprint feature with the reference voiceprint features. The second determination circuit determines whether a velocity of the sound source falls within a reference speed range according to a frequency variation of the sound signal that matches one of the voiceprint features. In response to the velocity of the sound source within the reference speed range, the control circuit controls the automatic door to be in an open state.

Abstract: Disclosed is an intelligent management system for a digital person dialogue based on Internet of Things. The system comprises a voice request question extraction module, configured to extract a voice request question input by a user and an input time point; a demand response time parsing module, configured to analyze a request urgency coefficient and a demand response time of the user; an adaptive response content search and determination module, configured to search and determine an adaptive response content; a dialogue scene recognition module, configured to recognize a dialogue scene; an adaptive dialogue intonation selection module, configured to select an adaptive dialogue intonation based on the dialogue scene; an adaptive dialogue language selection module, configured to determine an adaptive dialogue voice type; and a dialogue expression and implementation module, configured to perform a dialogue expression based on a dialogue platform.

Abstract: The present disclosure generally relates to decoding encoded files. More particularly, the present disclosure relates to systems and processes for determining the underlying language of encoded files having certain file formats with an undeterminable encoding. It may be difficult, time consuming, and costly to decode the files with undeterminable encodings as such files may require the language to be identified prior to decoding. Embodiments disclosed herein relate to determining the underlying language of the files by analyzing bit sequences in the encoded file and comparing the bit sequences against training data to determine matching patterns between the bit sequences in the file and the bit sequences known correspond to terms associated with a particular language.

Abstract: A computer system for extracting structured data from unstructured or semi-structured text in an electronic document, the system comprising: a graphical user interface configured to present to a user a graphical view of a document for use in training multiple data extraction models for the document, each data extraction model associated with a user defined question; a user input component configured to enable the user to highlight portions of the document; the system configured to present in association with each highlighted portion an interactive user entry object which presents a menu of question types to a user in a manner to enable the user to select one of the question types, and a field for receiving from the user a question identifier in the form of human readable text, wherein the question identifier and question type selected by the user are used for selecting a data extraction model, and wherein the highlighted portion of the document associated with the question identifier is used to train the sele

Abstract: A voice control method and apparatus for a device, a storage medium, and an electronic apparatus are provided. The method includes: acquiring a first voice feature of first voice data collected by a cleaning device, where the first voice data correspond to a first wake-up instruction sent by a use object, and the first wake-up instruction is used to wake up at least one of the cleaning device and a base station; acquiring a second voice feature of second voice data collected by the base station, where the second voice data correspond to the first wake-up instruction; and selecting a first device to be woken up from the cleaning device and the base station according to the first and second voice features, and waking up the first device, where the first device in a wake-up state is configured to respond to a voice instruction sent by the use object.

Abstract: A method uses natural language processing for visual analysis of a dataset. The method includes receiving a first natural language (NL) input directed to a data source, from a first client. The method also includes parsing the first NL input into tokens based on a grammar and the data source. The method also includes generating and outputting an intermediate NL response, to a second client, based on the tokens and output modality of the second client. In response to receiving a user input to provide missing information: the method also includes generating an input query based on the user input; and querying the data source using the input query, to obtain a result set. The method also includes generating and outputting, to the second client, a first NL output and a snapshot of a data visualization, based on the result set and the output modality of the second client.

Abstract: A system and method for creating accessibility of any website or application for people with sight, hearing or speech disabilities. The system and method can include receiving input of the website or the application to be accessed and an indicator as to specific disabilities a user, scoring the website or the application for its accessibility based on the specific disabilities of the user, and if the score is below a threshold, determining an alternative form for the input of the website or the application to accommodate the specific disabilities of the user.

Abstract: A method and electronic device for a vehicle are disclosed herein. The electronic device is mounted in the vehicle and includes a display, a memory storing voice commands, and a processor. The processor implements the method, including: obtaining at least one of vehicle driving information, occupant information, or display output information, generating one or more short commands by shortening one or more of the voice commands, based on the obtained at least one of the vehicle driving information, the occupant information and the display output information, and controlling the display to display one or more voice command guidance user interface (UI) displaying the one or more short commands.

Abstract: In various examples there is a computer-implemented method of database construction. The method comprises storing a knowledge graph comprising nodes connected by edges, each node representing a topic. Accessing a topic type hierarchy comprising a plurality of types of topics, the topic type hierarchy having been computed from a corpus of text documents. One or more text documents are accessed and the method involves labelling a plurality of the nodes with one or more labels, each label denoting a topic type from the topic type hierarchy, by, using a deep language model; or for an individual one of the nodes representing a given topic, searching the accessed text documents for matches to at least one template, the template being a sequence of words and containing the given topic and a placeholder for a topic type; and storing the knowledge graph comprising the plurality of labelled nodes.

Abstract: Apparatus (5) for assessing a quality of a communication session (3) between at least one first party (1) and at least one second party (2a, 2b . . . 2N), over a telecommunication network (4), comprising means for: monitoring (S1) said communication session by continuously receiving an audio stream associated with said communication session; converting (S2) language of said audio stream into text data; determining (S3), from said text data, at least first understandability quality features (UQFA, UQFG) and an information quality feature (IQF), said first understandability quality feature being representative of at least word articulation and grammar correctness within said language, and said information quality feature being representative of a comparison of the semantic content of said audio stream with a set of contents related to said audio stream; assessing (S4) said quality from said quality features.