Abstract: An automated content annotation workflow is disclosed. An example embodiment is configured for: registering a plurality of labelers to which annotation tasks are assigned; populating a labeling queue with content data to be annotated; assigning annotation tasks from the labeling queue to the plurality of labelers; enabling the plurality of labelers in an annotation review queue to modify or delete annotations applied by prior labelers; and evaluating a level of performance of the plurality of labelers in applying the annotations.

Abstract: An automated content labeling system is disclosed. An example embodiment is configured to: register a plurality of labelers to which annotation tasks are assigned; populate a labeling queue with content data to be annotated; assign annotation tasks from the labeling queue to the plurality of labelers; and provide a superpixel annotation tool enabling the plurality of labelers to configure a size of a segment cluster in an image of the content data, and select each segment cluster to be included in a segmentation feature with a specified object class, the segment clusters including similarly colored pixels from the image.

Abstract: An information processing apparatus includes a processor configured to display plural items in the order in which the items are to be processed, switch an item to be processed in accordance with a predetermined user operation, and output a process result that has been obtained before the user operation is performed in the case where the user operation is performed and an item that has been the item to be processed before the user operation is performed matches a predetermined output condition.

Abstract: Methods and systems are presented for dynamically generating programming code for implementing input data fields that automatically verifies formatting of input data based on platform agnostic formatting rules. A user interface page is obtained for display on a user device, wherein the user interface page is configured to obtain data from a user. Based on first programming code associated with the interface page, formatting rules are retrieved from a remote server. Second programming code that implements one or more input data fields on the user interface page is generated based on the formatting rules. The user interface page with the one or more input data fields is rendered on the user device for obtaining user data. Based on the second programming code, the user data inserted into the one or more input data fields is automatically formatted.

Abstract: Methods and systems for partial application data collection are disclosed. The method stores a network accessible application having a plurality of form fields. One or more processors generate a session identifier (ID) and timestamp for a user accessing the application. The one or more processors determine that a state within a form field of one of the plurality of form fields is in focus, the state of the form field being in focus when the user selects the form field. Additionally, it is determined when the state within the form field is no longer in focus. At that time, a countdown timer is initiated. When the countdown time reaches zero, any data entered into the form field is collected. The collected data and the session ID are then combined into a single user data file.

Abstract: This disclosure relates generally to visio-linguistic understanding. Conventional methods use contextual visio-linguistic reasoner for visio-linguistic understanding which requires more compute power and large amount of pre-training data. Embodiments of the present disclosure provide a method for visio-linguistic understanding using contextual language model reasoner. The method converts the visual information of an input image into a format that the contextual language model reasoner understands and accepts for a downstream task. The method utilizes the image captions and confidence score associated with the image captions along with a knowledge graph to obtain a combined input in a format compatible with the contextual language model reasoner. Contextual embeddings corresponding to the downstream task is obtained using the combined input. The disclosed method is used to solve several downstream tasks such as scene understanding, visual question answering, visual common-sense reasoning and so on.

Abstract: This disclosure describes methods, systems, and non-transitory computer readable media for automatically parsing infographics into segments corresponding to structured groups or lists and displaying the identified segments or reflowing the segments into various computing tasks. For example, the disclosed systems may utilize a novel infographic grouping taxonomy and annotation system to group elements within infographics. The disclosed systems can train and apply a machine-learning-detection model to generate infographic segments according to the infographic grouping taxonomy. By generating infographic segments, the disclosed systems can facilitate computing tasks, such as converting infographics into digital presentation graphics (e.g., slide carousels), reflow the infographic into query-and-response models, perform search functions, or other computational tasks.

Abstract: The present application discloses a method and apparatus for training a natural language processing model, a device and a storage medium, which relates to the natural language processing field based on artificial intelligence. An implementation includes: constructing training language material pairs of a coreference resolution task based on a preset language material set, wherein each training language material pair includes a positive sample and a negative sample; training the natural language processing model with the training language material pair to enable the natural language processing model to learn the capability of recognizing corresponding positive samples and negative samples; and training the natural language processing model with the positive samples of the training language material pairs to enable the natural language processing model to learn the capability of the coreference resolution task.

Abstract: A language correction system, a method therefor, and a language correction model learning method of the system are disclosed. The system comprises a correction model learning unit and a language correction unit. The correction model learning unit performs machine learning on a plurality of data sets consisting of ungrammatical sentence data and error-free grammatical sentence data respectively corresponding to the ungrammatical sentence data, so as to generate a correction mode for detecting grammatical sentence data corresponding to ungrammatical sentence data to be corrected. The language correction unit generates, for a sentence to be corrected, a corresponding corrected sentence by using the correction model generated by the correction model learning unit, and displays and outputs the corrected parts together with the generated corrected sentence.

Abstract: Systems and methods for identifying data strings in electronic documents using pattern recognition. The method includes receiving a first data string corresponding to an electronic reference document from a first database and a second data string corresponding to an electronic legal document from a second database. The method also includes processing the first data string into a first processed data string and processing the second data string into a second processed data string. The method also includes calculating a cosine similarity between the first processed data string and the second processed data string. The method also includes receiving a feedback score from a user which corresponds to an accuracy of the calculated cosine similarity. The method also includes calculating an adjusted cosine similarity between the first processed data string and the second processed data string based on the calculated cosine similarity and the feedback score.

Abstract: Provided are an item recall method and system, an electronic device and a readable storage medium. The item recall method includes: behavior data is acquired, where the behavior data includes items and item information of each item; target behavior data containing a retrieval category word is extracted from the behavior data; retrieval words of each item and a retrieval frequency of each retrieval word are acquired in a reverse correlation manner; word segmentation is performed on the item information to obtain multiple item segmented words; a similarity between all retrieval words and the multiple item segmented words is calculated; whether the similarity is greater than a first preset threshold or not is determined, and if yes, then a retrieval word is extracted as an expansion word of the retrieval category word; and item recall is performed according to the retrieval category word and the expansion word.

Abstract: Examples described herein generally relate to a computer system including a knowledge graph storing a plurality of entities. A mining of a set of enterprise source documents within an enterprise intranet is performed, by an enterprise named entity recognition (ENER) model, to determine a plurality of entity names. An entity record is generated within a knowledge graph for a mined entity name from the linked entity names based on an entity schema and ones of the set of enterprise source documents associated with the mined entity name. The entity record includes attributes aggregated from the ones of the set of enterprise source documents associated with the mined entity name.

Abstract: There is a need for more effective and efficient text categorization. This need can be addressed by, for example, techniques for semantic text categorization. In one example, a method includes determining an input vector-based representation of an input document; processing the input vector-based representation using a trained supervised machine learning model to generate the categorization based at least in part on the input vector-based representation, wherein: (i) the trained supervised machine learning model has been trained using automatically-generated training data, and (ii) the automatically generated training data is generated by determining an inferred semantic label for each unlabeled training document of one or more unlabeled training documents; and performing one or more categorization-based actions based at least in part on the categorization, and (iii) the labels are described by one or more short documents/short texts.

Abstract: A method is provided for creating a semantic model for submitting search queries thereto. The method includes an act of receiving data from one or more input sources in an entity and relationship capture service of a situational awareness engine. The method further includes an act of extracting entities and relationships between the entities in two or more extraction services, where the two or more extraction services include at least two of a table-to-graph service, an event-to-graph service, a sensor-to-graph service, a text-to-graph service, and an image-to-graph service. The method includes an act of generating a semantic model based on fusion and labeling the extracted data provided by the at least two extraction services, where the semantic model can receive a search query and respond to the search query based on the generated semantic model.

Abstract: Technical solutions relate to the natural language processing field based on artificial intelligence. According to an embodiment, a multilingual semantic representation model is trained using a plurality of training language materials represented in a plurality of languages respectively, such that the multilingual semantic representation model learns the semantic representation capability of each language; a corresponding mixed-language language material is generated for each of the plurality of training language materials, and the mixed-language language material includes language materials in at least two languages; and the multilingual semantic representation model is trained using each mixed-language language material and the corresponding training language material, such that the multilingual semantic representation model learns semantic alignment information of different languages.

Abstract: A multi-modal pre-training model acquisition method, an electronic device and a storage medium, which relate to the fields of deep learning and natural language processing, are disclosed. The method may include: determining, for each image-text pair as training data, to-be-processed fine-grained semantic word in the text; masking the to-be-processed fine-grained semantic words; and training the multi-modal pre-training model using the training data with the fine-grained semantic words masked.

Abstract: Implementations of the present specification provide a method and an apparatus for training a service model, and a method and an apparatus for determining a text classification category. During specific implementation, on the one hand, text is processed by using an encoding network to obtain a corresponding semantic vector; on the other hand, a relationship network is established for classification categories based on a hierarchical relationship, and the relationship network is processed by using a graph convolutional network, to fuse information of nodes to obtain category expression vectors. Then, the semantic vector of the text is fused with the category expression vectors to determine a prediction result of a classification category. In a phase of training a service model, the prediction result can be compared with a sample label to determine a loss and adjust model parameters.

Abstract: A method may include obtaining audio of a communication session between a first device of a first user and a second device of a second user. The method may further include obtaining a transcription of second speech of the second user. The method may also include identifying one or more first sound characteristics of first speech of the first user. The method may also include identifying one or more first words indicating a lack of understanding in the first speech. The method may further include determining an experienced emotion of the first user based on the one or more first sound characteristics. The method may also include determining an accuracy of the transcription of the second speech based on the experienced emotion and the one or more first words.

Abstract: A method for information processing in a user conversation, an electronic device and a storage medium thereof, which relate to the field of artificial intelligence and the natural language processing field, are disclosed. The method may include: acquiring conversation preceding information in a conversation between a first user and a second user; acquiring a target conversation strategy employed by the first user; generating initial reply content with a pre-trained response generating model according to the target conversation strategy and the conversation preceding information; and sending the initial reply content to a client of the first user, so as to display the initial reply content on a conversation interface of the first user with the second user.

Abstract: A computer-implemented method for predicting vehicle repairs based on a natural language processing (NLP). The method may include: receiving, by one or more processors, natural language data from a user interface; converting, by the one or more processors, the natural language data into text data; generating, by the one or more processors, natural language processed data (NLPD) by performing NLP on the text data and/or the natural language data; comparing, by the one or more processors, the NLPD to repair data stored in a repair database; determining, by the one or more processors, predicted repair data based on the comparing of the NLPD and the repair data; and transmitting, by the one or more processors, the predicted repair data to the user interface.

Abstract: An artificial intelligence (AI) model includes one or more feature models coupled to one or more observer models in a hierarchical fashion. The feature models are configured to process an input to detect different features within that input. The observer models are configured to analyze the operation of the feature models during processing of the input to generate various types of observations. One type of observation includes a natural language expression that conveys how various architectural and/or functional characteristics of a given feature model influence the processing of the input to detect features, thereby exposing the underlying mechanisms via which the given feature model operates.

Abstract: A system implements a dynamic correlated topic model (DCTM) to model an evolution of topic popularity, topic representation, and topic correlation within a set of documents, or other dataset, that spans a period of time. For example, the DCTM receives the set of documents and a quantity of topics for modeling. The DCTM processes the set by analyzing words of the documents, identifying word clusters representing the topics, and computing, for each topic, various distributions using continuous processes to capture a popularity, representation, and correlation with other topics across the period of time. In other examples, the dataset are user listening sessions comprised of media content items. Media content metadata (e.g., artist or genre) of the media content items, similar to words of a document, can be analyzed and clustered to represent topics for modeling by the DCTM.

Abstract: According to an embodiment, a control device for an image forming apparatus includes a controller that acquires reference information corresponding to expected signal strength values from a wireless tag on a sheet moving along a conveyance path and also acquires measured signal strength values for a first wireless tag on a sheet from a wireless tag communication device. The controller calculates a similarity value from the measured signal strength values and the expected signal strength values to determine whether the first wireless tag is on a sheet moving along the conveyance path. The controller controls the wireless tag communication device to write information to the first wireless tag when the calculated similarity value is above a threshold similarity value.

Abstract: Various implementations disclosed herein include multi-scale visual markers that convey information in multiple sets of markings using different respective appearance attributes. In some implementations, the appearance attribute of the markings of a first set of markings corresponds to a first encoding parameter and the appearance attribute of markings of a second set of markings corresponds to a second encoding parameter different from the first encoding parameter. In some implementations, the first set of markings encode first data and the second set of markings are different than the first set of markings and encode second data. In some implementations, the different appearance attributes are different scales (e.g., different sizes, different numbers of markings per unit of space, different contrast, different color characteristics, different wavelengths, different image sensor types, etc.).

Abstract: A fingerprint sensing apparatus is provided. A driving circuit drives a capacitive micromachined ultrasonic transducer (CMUT) array to emit a planar ultrasonic wave to a finger during a transmission period to generate reflected ultrasonic signals. CMUTs receive the reflected ultrasonic signals during a receiving period to generate sensing current signals. A sensing circuit senses the sensing current signals output by the CMUTs to generate fingerprint sensing signals.

Abstract: In a method for multipath reflection correction of acoustic signals received at an ultrasonic sensor, characteristics of multipath reflection signals of the ultrasonic sensor are accessed, wherein the characteristics of the multipath reflection signals include a relationship of primary signal contributions to multipath reflection signal contributions for acoustic signals received at the ultrasonic sensor at a plurality of times of flight for a plurality of locations of the ultrasonic sensor. Acoustic signals are received at the ultrasonic sensor over a time of flight range while a target is interacting with the ultrasonic sensor, wherein the acoustic signals include a primary signal contribution and a multipath reflection signal contribution. The characteristics of the multipath reflection signals are compared to received acoustic signals.

Abstract: A display substrate includes: a base substrate, a drive circuit on the base substrate, an insulating layer on a side of the drive circuit away from the base substrate, and a fingerprint recognition device on a side of the insulating layer away from the base substrate. The fingerprint recognition device includes: a first electrode, a photosensitive layer and a second electrode that are disposed in a stacked manner. The first electrode electrically connects to the drive circuit through a via hole running through the insulating layer; and the first electrode forms a recess in the via hole, and the recess is filled with an insulating material, so that a surface of the first electrode on a side away from the base substrate is a flat surface, and the photosensitive layer is disposed on the flat surface.

Abstract: The present disclosure provides a display apparatus and a method for manufacturing the same. The display apparatus includes: a self-luminous display module, a fingerprint identification module below the self-luminous display module, at least two light shield layers between the self-luminous display module and the fingerprint identification module, and a planarization layer correspondingly arranged on one side, facing the self-luminous display module, of each light shield layer. Each light shield layer includes a plurality of mutually-independent openings; openings of one of the at least two light shield layers and openings of another of the at least two light shield layers overlap in a direction perpendicular to the fingerprint identification module; the planarization layer between two adjacent light shield layers includes a groove; and the groove and the openings do not overlap in the direction perpendicular to the fingerprint identification module, and a light blocking part is arranged in the groove.

Abstract: A display device and a mobile terminal including the same includes a display area in which a first group of display pixels to which pixel data is written are arranged, and a sensing area in which a second group of display pixels and a plurality of sensor pixels are arranged; each of the display pixels includes a pixel circuit configured to drive a light emitting element; each of the sensor pixels includes a photosensor driving circuit configured to drive a photodiode; a low potential power voltage and a pixel driving voltage are applied to the pixel circuit and the photosensor driving circuit; a cathode electrode of the light emitting element and the cathode electrode of the photodiode share the same metal electrode, and are commonly connected to a low potential power line to which the low potential power voltage is applied.

Abstract: A display assembly and an electronic device are provided. The display assembly includes an optical sensor and a screen assembly which are stacked. The screen assembly is configured to emit a first optical signal toward one side away from the optical sensor. The optical sensor is configured to collect a fingerprint image of a target object according to a second optical signal obtained through reflection of the first optical signal. Components in the screen assembly, with light transmittance smaller than preset light transmittance, are at least partially stacked.

Abstract: A system for detecting and tracking an object that is exhibiting an anomalous behavior from a helicopter is disclosed. The system includes a search light connected to the helicopter; a camera; and a processor including an object detection module coupled to the search light and the camera, the object detection module being configured to receive a plurality of images from the camera, compare the plurality of images against a pattern database, determine the object is exhibiting the anomalous behavior and instruct the search light to point toward the object.

Abstract: A display device includes: a display panel to display an image by generating light, a sensor disposed under the display panel to convert a light energy into an electrical energy, a voltage generator connected to the sensor, and a capacitive element connected to the sensor. The electrical energy generated by the sensor is charged in the capacitive element. The sensor senses an object using the light reflected by the object.

Abstract: The present invention mainly provides an independent-operation intelligent image-tracking device without the need for a system for pairing and receiving at both ends. The structure of this invention includes a main body of the device and a single-chip unit combined in the main body. The main body of the device is connected with a dynamic mechanism, such as a motor with its transmission assembly, which can control the movement or rotation of the main body of the device. The single-chip unit is equipped with a lens-module, an image recognition module, an image analysis module, etc., which can obtain external images, calculate and analyze the image and shadow dynamics, and transmit the analysis results to the dynamic mechanism to perform corresponding dynamic displacements to achieve the purpose of automatically tracking target images.

Abstract: A three-dimensional position and posture recognition device speeds estimation of a position posture and a gripping coordinate posture of a gripping target product. The device includes: a sensor unit configured to measure a distance between an image of an object and the object; and a processing unit configured to calculate an object type included in the image, read model data of each object from the external memory, and create structured model data having a resolution set for each object from the model data, generate measurement point cloud data of a plurality of resolutions from information on a distance between an image of the object and the object, perform a K neighborhood point search using the structured model data and the measurement point cloud data, and perform three-dimensional position recognition processing of the object by rotation and translation estimation regarding a point obtained from the K neighborhood point search.

Abstract: A face authentication apparatus includes a face image acquisition unit that acquires a face image of an authentication target, a collation unit that performs face authentication by calculating similarity between face information of the face image of the authentication target and reference face image of each registered user and comparing the similarity with a threshold, a prediction unit that predicts a change in the similarity on the basis of similarity history on authentication success, and a threshold change unit that changes the threshold on the basis of the prediction result.

Abstract: Embodiments of the disclosure provide a method for classifying a face image, belonging to a field of artificial intelligence technologies. A face image is acquired. A face features is extracted from the face image. When a face feature collection does not meet a preset trigger condition, a category of the face feature is determined based on categories of existing face features in the face feature collection and a neighbor face feature algorithm. A category of the face image is determined based on the category of the face features.

Abstract: A dark circle detection and evaluation method includes: obtaining a to-be-processed image; extracting a dark circle region of interest from the to-be-processed image; performing color clustering on the dark circle region of interest to obtain n types of colors in the dark circle region of interest, where n is a positive integer; recognizing a dark circle region in the dark circle region of interest based on the n types of colors; and obtaining a dark circle evaluation result based on the dark circle region.

Abstract: Systems, and method and computer readable media that store instructions for.

Abstract: A driver recognition system of a vehicle identifies a person approaching the vehicle as a driver or other user. In one approach, the driver recognition system collects data using one or more cameras of the vehicle. The collected data corresponds to the person approaching the vehicle. Based on the collected data, a computing device determines (e.g., using a machine-learning model) whether the person is a user (e.g., driver) associated with the vehicle. If the person is a user associated with the vehicle, then the computing device causes one or more actions to be performed for the vehicle (e.g., controller configuration, boot up of a computing device, updating software using over-the-air update, etc.).

Abstract: A person detection device configured to analyze a fisheye image obtained by a fisheye camera installed above a to-be-detected area to detect a person existing in the to-be-detected area includes a human body detector configured to detect a human body candidate from a fisheye image and output, as a detection result, a bounding box indicating a region on the fisheye image of each human body candidate detected, a reference storage configured to prestore a reference for a shape and/or size of the bounding box for each position or area on the fisheye image, and a false detection determining unit configured to compare the shape and/or size of the bounding box of the human body candidate contained in the detection result with the reference corresponding to a position where the human body candidate is detected to determine whether the human body candidate results from false detection.

Abstract: An information processing apparatus (100) includes a collation unit (102) that collates first feature information extracted from a person included in a first image (20) with first registered feature information stored in a storage unit (110), and a registration unit (104) that stores, in a case where the first feature information is unable to be extracted from the person or a collation result in the collation unit (102) indicates a mismatch, second feature information extracted from the person in the storage unit (110) as second registered feature information, in which the collation unit (102) collates second feature information extracted from a person included in a second image (22) with the second registered feature information stored in the storage unit (110), and thus specifies the person corresponding to the second registered feature information in the second image (22).

Abstract: Systems, and method and computer readable media that store instructions for high-resolution face recognition.

Abstract: An image processing device includes one or more memories; and one or more processors coupled to the one or more memories and the one or more processors configured to store, in the one or more memories, a plurality of time-series images that has captured an object when the object is instructed to change an orientation of a face, extract a face region from each of the plurality of time-series images, obtain characteristic of changes of pixel values of a plurality of pixels arranged in a certain direction in the face region, specify an action of the object based on a time-series change in the characteristic of changes obtained from each of the plurality of time-series images, and determine authenticity of the object based on the action of the object.

Abstract: The present disclosure relates to an image processing method and apparatus. The method comprises: performing face and body feature extraction on an image to be processed to obtain image features including face features and/or body features, and the image to be processed comprises a first image and a second image; performing a face clustering operation, according to the face features extracted from the first image, to obtain a face clustering result; performing a body clustering operation on the second image to obtain a body clustering result, according to the face clustering result and the body features extracted from the first and second images, no face feature has been extracted from the second image; and obtaining, according to the face clustering result and the body clustering result, a clustering result for the image to be processed. The method can improve the recall rate while ensuring the accuracy of the clustering result.

Abstract: A security system includes a camera having a field of vision. The camera sends image frames to a machine readable memory. A processor is accessible to the machine readable memory. The processor performs acts of receiving from the camera a plurality of frames; detecting an object entering the field of vision; determining whether the object is a person; tracking a moving path of the person; and performing facial recognition on the person using at least one of the plurality of frames.

Abstract: Systems, and method and computer readable media that store instructions for high-resolution face recognition.

Abstract: An image processing method, an image processing device, a training method and a computer-readable storage medium. The image processing method includes: extracting a characteristic vector in an image to be recognized; based on the characteristic vector of the image to be recognized, acquiring a predicted score value of the image to be recognized; and based on the predicted score value, determining a category of an image information of the image to be recognized; wherein the image to be recognized is a face image, and the image information is a facial expression.

Abstract: A method to determine activity in a sequence of successively acquired images of a scene, comprises: acquiring the sequence of images; for each image in the sequence of images, forming a feature block of features extracted from the image and determining image specific information including a weighting for the image; normalizing the determined weightings to form a normalized weighting for each image in the sequence of images; for each image in the sequence of images, combining the associated normalized weighting and associated feature block to form a weighted feature block; passing a combination of the weighted feature blocks through a predictive module to determine an activity in the sequence of images; and outputting a result comprising the determined activity in the sequence of images.

Abstract: Methods for operating a motion recognition apparatus are disclosed. In some implementations, a method for recognizing a motion or gesture of an object may include operating an optical sensor device to capture light reflected from the object under illumination by light emitted toward the object, generating, by comparing the emitted light to the reflected light, a depth image including distance information indicating a distance between the optical sensor device and the object, generating, based on the light reflected from the object, an infrared image including infrared image information associated with the object, and determining the motion of the object based on at least one of the depth image and the infrared image.

Abstract: An electronic device includes memory circuitry, and processor circuitry having an action detection circuitry configured to operate according to an action detection model for detecting an action based on a machine-learning scheme. The processor circuitry being configured to obtain sensor data; generate, based on the sensor data, a set of features associated with a frame; determine, based on the set, using the action detection model, whether the frame corresponds to a sub-action; apply a nondeterministic finite automaton, NFA, scheme, to the determined sub-action for the frame, wherein the NFA scheme has a set of states associated with corresponding sub-actions and is configured to output one or more action classes; determine, using the NFA scheme, an action class; detect the action based on the action class; and increment an action counter based on the detected action.