Abstract: Techniques for identifying vocabulary associated with semantic objects used for generating natural language text with a natural language generation (NLG) system, the semantic objects including a first semantic object having a first set of ordered attributes. The techniques include: obtaining text segments; identifying, from among the text segments and using at least one first machine learning classifier, groups of text segments corresponding to respective semantic objects in the plurality of semantic objects; identifying, from the groups of text segments and using at least one second machine learning classifier, a plurality of vocabularies for the plurality of semantic objects; and generating natural language text using the NLG system, the plurality of vocabularies, and the plurality of semantic objects; and outputting the generated natural language text.

Abstract: A system and method automatically associates an image and passage from a text. In the system and method, a user can choose or supply an image and the system and/or method will choose a limited selection of relevant word passages for the image from a relatively large volume of potential passages. The system and method utilize a computer system wherein a concept generator and a passage generator processes the content of the image so as to assign words to describe the content, then weight the descriptive words (tags) and assign passages based on the tags and weighting. The passages can be filtered so as to remove inappropriate passages.

Abstract: System that facilitates rapid onboarding of an autonomous (cashier-less) store by capturing images of the store's items from multiple angles, with varying backgrounds, and that builds a classifier training dataset from these images. The system may have cameras in different positions, and backgrounds that generate different background colors. It may have a platform for items that can be switched between a transparent and non-transparent state; cameras below the platform may therefore capture images of the bottom side of the item when the platform is transparent. When an item is placed in the imaging system, a fully automated process may generate a sequence of background colors and may capture and process images from all of the cameras to create training images. Images of the item from multiple angles, including views of the entire external surface of the item, may be captured without requiring an operator to move or reorient the item.

Abstract: An artificial intelligence apparatus for recognizing an object included in image data can include a camera, a communication modem, a memory configured to store an image recognition model, a natural language processing (NLP) model, and an NLP model-based image recognition model learned based on the NLP model, and a processor is configured to receive image data from the camera or the communication modem, in response to recognizing an object included in the image data using the image recognition model, generate first recognition information on the object included in the image data, and in response to the recognizing the object included in the image data using the image recognition model being unsuccessful, generate second recognition information on the object included in the image data based on recognizing the object using the NLP model-based image recognition model.

Abstract: This application discloses a prosodic pause prediction method, a prosodic pause prediction device and an electronic device. The specific implementation scheme includes: obtaining a first matrix by mapping a to-be-tested text sequence through a trained embedding layer, where the to-be-tested text sequence includes a to-be-tested input text and an identity of a to-be-tested speaker; inputting the first matrix into a trained attention model, and determining a semantic representation matrix by the trained attention model; and, performing prosodic pause prediction based on the semantic representation matrix and outputting a prosodic pause prediction result of each word in the to-be-tested input text.

Abstract: A method of detecting an edge of a support surface in an imaging controller includes: obtaining image data captured by an image sensor and a plurality of depth measurements captured by a depth sensor, the image data and the plurality of depth measurements corresponding to an area containing the support surface; detecting preliminary edges in the image data; applying a Hough transform to the preliminary edges to determine Hough lines representing candidate edges of the support surface; segmenting the plurality of depth measurements to assign classes to each pixel, each class defined by one of a plurality of seed pixels, wherein the plurality of seed pixels are identified from the depth measurements based on the Hough lines; and detecting the edge of the support surface by selecting a class of pixels and applying a line-fitting model to the selected class to obtain an estimated edge of the support surface.

Abstract: Methods and systems for processing an image are described. A saliency map is generated from the image. The saliency map indicates one or more salient portions of the image that have saliency values satisfying a saliency criterion. A scene graph is generated for at least the one or more salient portions. The scene graph represents a plurality of objects detected in the image. The scene graph further represents one or more relationships between each respective object pairs. One or more dataset entries associated with the image are generated. Each of the one or more relationships for each of the one or more object pairs is indicated by a respective dataset entry. The one or more dataset entries are stored in a first dataset.

Abstract: A vision-based system for identifying textual sentiment present in a text source image, comprising a camera module adapted to capture the text source image, and a vision-based text analyzing device adapted to analyze the input image, identify a text input string, and determine the sentiment embodied by the text input string, the vision-based text analyzing device is further adapted to revise the text input string to embody an alternate sentiment, the vision-based system has an output display for presenting sentiment classification results and the revised text string via in an augmented reality format.

Abstract: A method may include presenting a user interface on a computing device, the user interface including a character input portion and a predictive suggestion portion; converting, using at least one processor, a set of characters entered in the character input portion into a word vector; inputting the word vector into a neural network; based on the inputting, determining a set of output terms from the neural network; querying a data store to retrieve a user specific data value for an output term of the set of output terms; and presenting a paired output term that includes the output term and the user specific data value in the predictive suggestion portion of the user interface.

Abstract: An information processing apparatus comprises an extraction unit configured to extract, as an image edge group, a plurality of image edges in the vicinity, which are detected from a captured image obtained by capturing a target object; a determination unit configured to determine a positional relationship between the image edges included in the image edge group; a holding unit configured to hold a positional relationship between model edges that constitute a three-dimensional model of the target object; a selection unit configured to select a combination of an image edge and a model edge, which correspond to each other, based on matching between the positional relationship of the image edges and the positional relationship of the model edges; and an collation unit configured to estimate a position/orientation of the target object based on the combination.

Abstract: A method modifies a social media post in response to determining that a hashtag for the social media post diverges from how the hashtag is used by other social media posts. A computing device receives a social media post, which contains a hashtag, and analyzes the social media post using natural language processing and a sentiment analysis to determine a sense and sentiment of the social media post. The computing device receives a plurality of reference social media posts, where each of the reference social media posts contains the hashtag, and utilizes the natural language processing and the sentiment analysis to determine a sense and sentiment of the reference social media posts. In response to determining that the sense and sentiment of the social media post do not match the sense and sentiment of the reference social media posts, the computing device causes the social media post to be altered.

Abstract: A method and an apparatus for processing word vectors of a neural machine translation model, and a non-transitory computer-readable recording medium are provided. In the method, word vectors that are input to an encoder and a decoder of a neural machine translation model are updated using semantic information among head representations at the same time and semantic information among head representations at different times, and the model is trained or translation is performed using the updated word vectors, thereby improving the model performance of the neural machine translation model.

Abstract: Embodiments of this application provide a text recommendation method performed at an electronic device. The method includes: extracting feature content of from the a target text; processing the feature content by using at least two text analysis models to obtain at least two semantic vectors; integrating the at least two semantic vectors into an integrated semantic vector of the target text; selecting, according to the integrated semantic vector and an integrated semantic vector of at least one to-be-recommended text, a recommended text corresponding to the target text from the at least one to-be-recommended text. Because the integrated semantic vector of the target text is obtained based on the at least two text analysis models, the integrated semantic vector has a stronger representing capability. When text recommendation is subsequently performed, an association degree between the recommended text and the target text can be increased, thereby improving recommendation accuracy.

Abstract: An apparatus and method for generating a dialogue state tracking model. The apparatus retrieves a field feature corresponding to a queried field from a database according to the queried field corresponding to a queried message. The apparatus retrieves a candidate-term feature corresponding to each of at least one candidate-term corresponding to the queried field from the database, and integrates them into an integrated feature. The apparatus also generates at least one relation sub-sentence of a reply message corresponding to the queried message and generates a sentence relation feature according to the at least one relation sub-sentence. The apparatus further generates a queried field related feature according to the field feature, the integrated feature and the sentence relation feature and trains the dialogue state tracking model according to the queried field related feature.

Abstract: Affected areas of images captured by cameras that are experiencing one or more anomalies may be modified by inpainting portions of counterpart images onto the affected areas. Where a camera provided aboard an operating vehicle (e.g., an aerial vehicle) is experiencing an anomaly that adversely affects the quality of images captured thereby, counterpart images that are captured by other cameras (e.g., cameras provided aboard the vehicle or elsewhere) at the same time, or by the affected camera at an earlier time, may be used for inpainting. Portions of counterpart images may be selected for inpainting according to a transform function that takes into account a pose of an affected camera, dynamics of an operating vehicle or any other factors. Images that have been modified by inpainting may be utilized by an operating vehicle for any purpose, such as computer vision applications, navigation, guidance, or collision avoidance.

Abstract: A data management server computer (“server”) and related methods are disclosed to enhance electronic documents or search processes related to skill data, including skill names, using machine learning techniques. The server is programmed to organize skill names from a dictionary into a knowledge graph. The knowledge graph includes nodes that represent the skill names and words of the skill names and edges that represent syntactic or semantic relationships. The server is programmed to respond to skill-related requests using the knowledge graph and track access patterns of the knowledge graph in responding to the requests. The replies to the request can be generated based on the structure or access patterns of the knowledge graph and used as search results or skill-related data, such as resumes or job descriptions.

Abstract: Structured and semi-structured databases and files are processed using natural language processing techniques to impute data for null value tokens in database records from other records that have non-null values for the same attributes. Vector embedding techniques are used, including, in some cases, appropriately tagging null value tokens to reduce or eliminate their undue impact on semantic vectors generating using a neural network.

Abstract: A computer system evaluating an input segment of a communication, in parallel, by a baseline classification model trained with baseline passages indicating dictionary meaning and multiple generative sequence models each trained to classify a particular passage from among multiple indirect passages indicating usage with an indirect meaning, to receive a separate score from the baseline classification model and each of the generative sequence models, each separate score indicating a classification probability for the input segment. The computer system, responsive to one or more particular scores generated by one or more of the generative sequence models exceeding a baseline score generated by the baseline classification model summed with a tuning factor, flagging the input segment as having a potentially indirect meaning.

Abstract: Aspects described herein include a method of semantic parsing, and related system and computer program product. The method comprises receiving an input comprising a plurality of words, generating a structured representation of the plurality of words, encoding the structured representation into a latent embedding space, and decoding the encoded structured representation from the latent embedding space into a logical representation of the plurality of words.

Abstract: An artificial neural network (ANN) determines a conversation snippet sentiment score based on content of the conversation snippet and contextual attributes associated with the conversation snippet. Contextual attributes may include, for example, a role within an organizational hierarchy of a user participating in the conversation snippet. Information representing the content is input into a hidden layer sequence of the ANN; information representing the contextual attributes is input into another hidden layer sequence of the ANN. Additionally or alternatively, a weighing engine determines a topical sentiment score by aggregating weighted conversation snippet sentiment scores.