Abstract: An apparatus for matching video records with postings using audiovisual data processing is provided. Apparatus may include at least a processor and a memory communicatively connected to the processor. The memory may contain instructions configuring the at least a processor to receive a posting including a plurality of criteria, receive a video record, extract a plurality of textual elements from the video record, identify a plurality of credentials from the video record, generate a compatibility score for the video record based on the plurality of criteria and the plurality of credentials, and match the video record with the posting based on the compatibility score.

Abstract: A response device includes a data acquirer configured to acquire data to which metadata is added, a database register configured to generate a tag on the basis of the metadata and register the generated tag in a database in association with the data to which the metadata is added, an utterance content interpreter configured to interpret content of an utterance of a user, a searcher configured to search the database using the tag included in the utterance content when an intention to search for the data has been interpreted by the utterance content interpreter, and a responder configured to cause an outputter to output information according to a search result of the searcher.

Abstract: An information processing system including a video device for displaying a video and a wireless terminal for communicating with the video device and communicating with other apparatuses via a network, wherein the video device transmits video information captured by a camera and sound information collected by a microphone when reception from the wireless terminal of information pertaining to a video to be displayed is detected, and outputs the received sound information using a speaker. The wireless terminal transmits information pertaining to the video to be displayed to the video device, the other apparatuses receive the captured video information and collected sound information via the network, and transmit information pertaining to the video to be displayed and information pertaining to the sound to be outputted.

Abstract: An off-policy reinforcement learning actor-critic neural network system configured to select actions from a continuous action space to be performed by an agent interacting with an environment to perform a task. An observation defines environment state data and reward data. The system has an actor neural network which learns a policy function mapping the state data to action data. A critic neural network learns an action-value (Q) function. A replay buffer stores tuples of the state data, the action data, the reward data and new state data. The replay buffer also includes demonstration transition data comprising a set of the tuples from a demonstration of the task within the environment. The neural network system is configured to train the actor neural network and the critic neural network off-policy using stored tuples from the replay buffer comprising tuples both from operation of the system and from the demonstration transition data.

Abstract: Speech is digitized and analyzed by a speech-recognition platform to produce raw text sentences. In various embodiments, the recognized words of each sentence are tokenized based on a grammar, which may be selected by a Recognition Context Controller (RCC) using a context database. A Medical Context Semantic Library (MCSL) contains all medically relevant terms recognized by the system and, once the grammar is selected, the MCSL is used to select a semantic template (consisting of one or more hierarchically organized data structures whose root is a “Concept”). Recognized words are mapped to tokens based on the operative grammar to fill the Concept tree(s). The grammar and the Concept trees can potentially shift after each sentence based on the RCC's analysis. The trees accumulate and are filled as sentences are analyzed. Once all of the sentences have been analyzed, the trees have been filled to the extent possible. Concepts may be organized into higher-level Observations.

Abstract: An information processing apparatus includes a memory storing instructions. The instructions cause the apparatus to extract a plurality of local features from data indicating a speech, the characteristics of feature extraction being formed through learning; and to encode a series of chronological features of the data based on the plurality of local features, characteristics of encoding the series of chronological features being formed through learning. The instructions also cause the apparatus to generate information obtained by weighting features at a specific point in time associated with emotion classification, of the series of chronological features encoded, characteristics of weighting the features at the specific point in time being formed through learning; and to classify emotion corresponding to the data using the information obtained by weighting the features at the specific point in time, characteristics of classification being formed through learning.

Abstract: In one embodiment, a method includes accessing visual data from a client system associated with a user, wherein the visual data comprises images portraying one or more objects, receiving, from the client system, a user request, wherein the user request comprises a coreference to a target object, resolving the coreference to the target object from among the one or more objects, resolving the target object to a specific entity, and sending, to the client system, instructions for providing a response to the user request, wherein the response comprises attribute information about the specific entity.

Abstract: A terminal, a method for voice control, and related products are provided. The method for voice control is applicable to a terminal including a LPWAN communication module, and the method includes the following. A first enhanced LPWAN with at least one smart device is established. Voice data of a user is obtained and a control intention of the user for the at least one smart device is determined according to the voice data. A dedicated control instruction for a first smart device of the at least one smart device is determined according to the control intention, where the dedicated control instruction is used to instruct the first smart device to perform a corresponding operation. The dedicated control instruction is sent to the first smart device.

Abstract: In one embodiment, the disclosed technology involves: digitally generating and storing a machine learning statistical topic model in computer memory, the topic model being programmed to model call transcript data representing words spoken on a call as a function of one or more topics of a set of topics that includes pre-seeded topics and non-pre-seeded topics; programmatically pre-seeding the topic model with a set of keyword groups; programmatically training the topic model using unlabeled training data; conjoining a classifier to the topic model to create a classifier model; programmatically training the classifier model using labeled training data; receiving target call transcript data; programmatically determining at least one of one or more topics of the target call or one or more classifications of the target call; and digitally storing the target call transcript data with additional data indicating the determined topics and/or classifications of the target call.

Abstract: A method includes selecting a customer of a company; constructing a digital footprint of the selected customer. The method includes inputting the digital footprint to an artificial intelligence (AI) engine. The method includes obtaining one or more probability values from the AI engine based on the input digital footprint. The method includes selecting a call driver, from among a plurality of call drivers, as a predicted call driver. The method includes providing the predicted call driver to a call center associated with the company.

Abstract: The invention discloses a method for predicting a residual strength of a composite material after impact. The method includes: acquiring a first frequency value of a composite material to be tested after impact damage; determining a first frequency change rate according to an initial frequency value and the first frequency value of the composite material to be tested; inputting the first frequency change rate to a pre-constructed residual strength prediction model to predict a first residual strength of the composite material to be tested; the first frequency value and the initial frequency value being obtained by a modal test. Further, residual strength prediction may be performed on the composite material in service, the position and the size of impact damage, the impact energy, the shape and quality of the impact object are not required to be determined, which avoids a rigidity test of the composite material in service.

Abstract: A virtual data lake system created with browser-based decentralized data access and analysis is disclosed herein. As contemplated by the present disclosure, the system provides a single interface that allows a user to access and analyze multiple enterprise data storage locations remotely and simultaneously while presenting and reporting information from the multiple sources in a single, uniform display. Such a solution allows a user to analyze and cross-reference data stored in multiple locations in real time without requiring the actual data files to be displaced or combined. The system further implements interactive artificial intelligence, natural language processing, and workflow-based operations for improved user access and functionality.

Abstract: [Problem] There are proposed an information processing apparatus, an information processing method, and a program, which are capable of learning a meaning corresponding to a speech recognition result of a first speech adaptively to a determination result as to whether or not a second speech is a restatement of the first speech. [Solution] An information processing apparatus including: a learning unit configured to learn, based on a determination result as to whether or not a second speech collected at second timing after first timing is a restatement of a first speech collected at the first timing, a meaning corresponding to a speech recognition result of the first speech.

Abstract: A system is provided for a self-learning policy engine that can be used by various spoken language understanding (SLU) processing components. The system also provides for sharing contextual information from processing performed by an upstream SLU component to a downstream SLU component to facilitate decision making by the downstream SLU component. The system also provides for a SLU component to select from a variety of actions to take. A SLU component may implement an instance of the self-learning policy that is specifically configured for the particular SLU component.

Abstract: The disclosure herein describes a system and method for attentive sentence similarity scoring. A distilled sentence embedding (DSE) language model is trained by decoupling a transformer language model using knowledge distillation. The trained DSE language model calculates sentence embeddings for a plurality of candidate sentences for sentence similarity comparisons. An embedding component associated with the trained DSE language model generates a plurality of candidate sentence representations representing each candidate sentence in the plurality of candidate sentences which are stored for use in analyzing input sentences associated with queries or searches. A representation is created for the selected sentence. This selected sentence representation is used with the plurality of candidate sentence representations to create a similarity score for each candidate sentence-selected sentence pair.

Abstract: The present disclosure relates to techniques for automatically training a chatbot using utterances not understood by the chatbot itself and thus improve its understanding in a fast, effective and unsupervised way. This disclosure implements techniques to update known intents, to create new ones, and to modify the dialog manager so that new responses can be provided. Known intents can be updated with the chatbot's non-understood messages in three different ways: 1) when the user is asked to reformulate his request due to the chatbot not understanding the user, and the confidence of the new intent is greater than a confidence threshold, and the new detected intent is the same as the previous intent; 2) when the user accepts an intent suggestion from the chatbot; and 3) when the agent accepts a response suggestion from the chatbot. New intents, on the other hand, can be automatically created and automatically associated with real answers provided by human agents.

Abstract: An emotion estimation system includes a feature amount extraction unit, a vowel section specification unit, and an estimation unit. The feature amount extraction unit analyzes recorded produced speech to extract a predetermined feature amount. The vowel section specification unit specifies, based on the feature amount extracted by the feature amount extraction unit, a section in which a vowel is produced. The estimation unit estimates, based on the feature amount in a vowel section specified by the vowel section specification unit, an emotion of a speaker.

Abstract: A camera includes a first microphone, a second microphone, and a processor. The processor determines a correlation metric between portions of audio signals obtained from the first and second microphones. The camera includes a memory to store the portions of the audio signals as portions of an output audio signal.

Abstract: Techniques for creating and calibrating natural-language understanding (NLU) machine learning models are described. In certain embodiments, a training service tunes parameters of a function, taking the output from an NLU machine learning model as an input of the function, to calibrate the NLU machine learning model's output to optimize the interpretability of the resulting output, e.g., confidence score(s). Embodiments herein include generating, by the NLU machine learning model, an output based at least in part on an input (e.g., utterance) from a user, and applying a tuned, output modifying function to the output from the NLU machine learning model to generate a modified output. An inference may be generated based at least in part on the modified output.

Abstract: Disclosed herein are system, method, and computer program product embodiments for categorizing customer complaints on social media using a model trained on customer voice calls or chats with agents. Additionally, users interested in monitoring regulatory compliance issues based on customer complaints can receive notifications regarding complaints that are linked to regulatory topic areas, without the need to manually scan vast numbers of social media postings.