| 05th week of 2021 patent applcation highlights part 51 |
| Patent application number | Title | Published |
|---|
| 20210034875 | COMPUTER IMPLEMENTED STRUCTURAL THERMAL AUDIT SYSTEMS AND METHODS - In one preferred form of the present invention, there is provided a computer implemented thermal audit method ( | 2021-02-04 |
| 20210034876 | VIDEO PRIVACY USING MACHINE LEARNING - A method, system and computer program product for providing video privacy is provided herein. First video data captured by a video camera is received. A context for the first video data is determined. It is determined that the context matches a privacy context from a set of privacy contexts identified using machine learning. In response to the context matching the privacy context, at least a portion of second video data is blocked that is captured by the video camera subsequent to the first video data. | 2021-02-04 |
| 20210034877 | METHOD FOR IDENTIFYING A PERSON IN A VIDEO, BY A VISUAL SIGNATURE FROM THAT PERSON, ASSOCIATED COMPUTER PROGRAM AND DEVICE - A method includes for each of a plurality of successive images in a camera video stream, searching for at least one person present in the image and defining, in the image, for each person found, a field, called person field, at least partially surrounding that person; for each of at least one person found, gathering into a track segment several person fields derived from successive images and at least partially surrounding that same person; for each track segment, identifying the person in that track segment, by a visual signature from that person, this identification including: for each person field in the track segment, determining a visual signature from the person in that track segment, called local visual signature; determining an aggregated visual signature from the local visual signatures; and identifying the person in that track segment from the aggregated visual signature. | 2021-02-04 |
| 20210034878 | Image Detection Using Multiple Detection Processes - In some embodiments, a first detector generates a first output based on a first probability that an image was inserted in a video. The first detector is trained with a set of known images to detect the set of known images. A second detector generates a second output based on a second probability that an image was inserted in the video. The second detector is used to detect the set of unknown images without training. The method analyzes the first output from the first detector based on the probability of the image existing in the video and the second output from the second detector based on the probability of the image existing in the video to generate a combined score from the first output and the second output. An indication of whether the image is detected in the video is output based on the combined score. | 2021-02-04 |
| 20210034879 | System And Method for Identifying Cognate Image Sequences - A system and method of identifying cognate image sequences is disclosed that examines significant frames of a stream of video images using an array of image indexes. The image index array includes image indexes obtained by at least two different image indexing methods. These are compared to a corresponding array of image indices of significant frames of known image sequences. An image quality indicator is used to determine which set of image index thresholds to use in making the comparison. These thresholds are more stringent for higher quality frames. Two image sequences are considered cognate when a string of sufficiently many sequential frame matches is established. In an alternate embodiment, image blurriness is also, or instead, used to determine the appropriate set of image index thresholds. The sets of image index thresholds are determined using machine learning on a curated set of representative images. | 2021-02-04 |
| 20210034880 | TIME-BASED AUTOMATIC VIDEO FEED SELECTION FOR A DIGITAL VIDEO PRODUCTION SYSTEM - A video production device is deployed to produce a video production stream of an event occurring within an environment that includes a plurality of different video capture devices capturing respective video input streams of the event. The video production device is programmed and operated to: receive a plurality of video input streams from the plurality of different video capture devices; automatically sequence through the streams to select one of them as a current video output stream, in accordance with a predetermined switching sequence associated with the video capture devices; and provide the selected stream as the current video output stream for a designated time interval associated with the video capture device that captured the selected video input stream. | 2021-02-04 |
| 20210034881 | MONITORING METHOD, APPARATUS AND SYSTEM, ELECTRONIC DEVICE, AND COMPUTER READABLE STORAGE MEDIUM - Embodiments of a surveillance method, apparatus, system, electronic device and computer-readable storage medium are provided. In the surveillance method, a non-visible light image and a target image are obtained, wherein the target image is generated from a visible light signal captured during a capture period of the non-visible light image. The method then detects whether an object is present in the non-visible light image. When an object is detected in the non-visible light image, a second location area of the object in a visible light image is determined according to a first location area of the object in the non-visible light image, such that surveillance of the object is implemented based on the visible light image, wherein the visible light image is an image determined based on the target image. Compared to the relevant art, the embodiments proposes to use the result of the object detection performed on the non-visible light image to determine the result of the object detection performed for the visible light image corresponding to the non-visible light image, which guarantees that the object detection result of the visible light image also has a high accuracy and thereby guarantees the effect of intelligent surveillance. | 2021-02-04 |
| 20210034882 | DOOR LOCK SYSTEM WITH ONE OR MORE VIRTUAL FENCES - An intelligent door lock system is coupled to a door at a dwelling. A sensor is at the dwelling. The sensor is coupled to a drive shaft of a lock device to assist in locking and unlocking a lock of a lock device at the door. The lock device is coupled to the sensor and includes a bolt. An engine, an energy source and a memory are coupled together. A camera is coupled to or part of the intelligent door lock system. The camera is configured to define a safe zone in the dwelling in which an occupant, and a non dwelling occupant third person is allowed into the dwelling. | 2021-02-04 |
| 20210034883 | METHODS AND APPARATUS TO COUNT PEOPLE IN IMAGES - Example apparatus disclosed herein include a memory and a processor to execute instructions to identify a first set of face rectangles and a second set of face rectangles in a frame pair of image data corresponding to a media environment, the first set of face rectangles corresponding to a first image sensor and the second set of face rectangles corresponding to a second image sensor, remove first face rectangles from the first set of face rectangles and the second set of face rectangles when the first face rectangles are determined to correspond to false positive face detections, group second face rectangles that remain in the first set of face rectangles and the second set of face rectangles after removal of the first face rectangles to form groups of face rectangles, and generate a count of people identified in the media environment based on a number of the groups. | 2021-02-04 |
| 20210034884 | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, PROGRAM, AND MOBILE OBJECT - An information processing apparatus according to an embodiment of the present technology includes a first acquisition unit, a second acquisition unit, and a generation unit. The first acquisition unit acquires peripheral information regarding a periphery of a first mobile object. The second acquisition unit acquires, from an apparatus different from the first mobile object, attribute information regarding an attribute of a second mobile object present in the periphery of the first mobile object. The generation unit generates learning data for extracting an attribute of a target mobile object, on the basis of the acquired peripheral information and the acquired attribute information. | 2021-02-04 |
| 20210034885 | Automated Capture Of Image Data For Points Of Interest - Automated capture of image data for points of interest may be implemented for points of interest in an environment external to a vehicle. Sensors implemented as part of a vehicle may collect sensor data for an environment. Processing of the sensor data may be performed to detect points of interest in the environment. In response to detecting a point of interest, image data may be captured by one or more of the sensors implemented at the vehicle. Different types of image data may be captured, such as panoramic images and three-dimensional reconstructions of a scene. Metadata may be generated for captured image data which may describe the point of interest that is captured by the image data. The image data and the metadata may be stored locally at the vehicle or to a remote data store. The image data may also be shared with other computing devices. | 2021-02-04 |
| 20210034886 | SYSTEM AND METHOD FOR PREVENTING SCRAPES ON VEHICLE CARRIAGE - Methods and systems for identifying geographic locations where a vehicle carriage of a vehicle may be scraped. The system includes a transceiver of the vehicle configured to receive scraping location data indicating one or more scraping locations, each scraping location being a geographic location where a risk of scraping the vehicle carriage exists. The system also includes a location sensor configured to detect a current location of the vehicle. The system also includes an electronic control unit (ECU) of the vehicle configured to determine whether the vehicle is approaching a scraping location of the one or more scraping locations based on the current location of the vehicle and the scraping location data. The system also includes an input/output device configured to provide a notification that the vehicle is approaching the scraping location of the one or more scraping locations. | 2021-02-04 |
| 20210034887 | PASSENGER STATE DETECTION DEVICE, PASSENGER STATE DETECTION SYSTEM, AND PASSENGER STATE DETECTION METHOD - A passenger state detection device ( | 2021-02-04 |
| 20210034888 | Method and System for Vision-Based Vehicle Interior Environment Sensing Guided by Vehicle Prior Information - A method for operating a vehicle including a vehicle sensing system includes generating a baseline image model of an cabin of the vehicle based on image data of the cabin of the vehicle generated by an imaging device of the vehicle sensing system, the baseline image model generated before a passenger event, and generating an event image model of the cabin of the vehicle based on image data of the cabin of the vehicle generated by the imaging device, the event image model generated after the passenger event. The method further includes identifying image deviations by comparing the event image model to the baseline image model with a controller of the vehicle sensing system, the image deviations corresponding to differences in the cabin of the vehicle from before the passenger event to after the passenger event, and operating the vehicle based on the identified image deviations. | 2021-02-04 |
| 20210034889 | SYSTEM AND METHOD TO DETECT DRIVER INTENT AND EMPLOY SAFE DRIVING ACTIONS - Embodiments are directed towards the interpretation of driver intent and communication of that intent with autonomous vehicles at a traffic intersection. A computing device that sits on the dashboard of a vehicle includes at least one camera, an output device, and circuitry. The computing device captures first images of the driver in the vehicle and second images of the traffic intersection. The computing device identifies another vehicle at or approaching the intersection based on an analysis of the second images. The computing device determines an attention direction and hand movement of the driver based on an analysis of the first images to determine a driving intent of the driver. The computing device provides information to the other vehicle indicating the driving intent of the driver. The computing device may also obtain the driving intent of the other vehicle and provide information to the driver indicating the other vehicle's intent. | 2021-02-04 |
| 20210034890 | DRIVER MONITORING APPARATUS AND METHOD - The present disclosure relates to an apparatus and method for monitoring a driver. The apparatus includes one or more sensors that are embedded in a seat belt and that measure electrical resistances according to a movement of the driver, respectively, and a controller that estimates a posture of the driver, based on changes of the electrical resistances measured by the sensors. | 2021-02-04 |
| 20210034891 | EMOTION ESTIMATION DEVICE - An emotion estimation device includes: an emotion estimation unit estimating a user emotion as an emotion associated with an index value of a facial expression of a user between (i.e., involving/during a period having) a plurality of consecutive events, based on a correspondence between (A) the index value which is at least one of (i) a facial expression score of (i.e., representing) the facial expression of the user and (ii) a feature amount about a change of such facial expression score and (B) the facial expression of the user correlated with the plurality of consecutive events. | 2021-02-04 |
| 20210034892 | Detecting Spoofing Talker in a Videoconference - A videoconferencing device at an endpoint determines whether a person is a real person standing in front of a display device or if the person is instead an image rendered by a display device. In the first instance the real person will be included in a video feed for transmission to a remote endpoint. In the second instance, images of the display device on which the person is rendered will not be included in the video feed. | 2021-02-04 |
| 20210034893 | SPOOFING DETECTION APPARATUS, SPOOFING DETECTION METHOD, AND COMPUTER-READABLE RECORDING MEDIUM - A spoofing detection apparatus comprises obtaining, from an image capture apparatus, a first image frame that includes the face of a subject person obtained when a light-emitting apparatus is emitting light and a second image frame that includes the face of the subject person obtained when the light-emitting apparatus is turned off, extracting, from the first image frame, information specifying a face portion of the subject person, and extract, from the second image frame, information specifying a face portion of the subject person, extracting a portion that includes a bright point formed by reflection in an iris region of an eye of the subject person, from the first image frame, extracts a portion corresponding to the portion that includes the bright point, from the second image frame, and calculates a feature that is independent of the position of the bright point, and determining authenticity of subject person based on the feature. | 2021-02-04 |
| 20210034894 | SPOOFING DETECTION APPARATUS, SPOOFING DETECTION METHOD, AND COMPUTER-READABLE RECORDING MEDIUM - A spoofing detection apparatus | 2021-02-04 |
| 20210034895 | MATCHER BASED ANTI-SPOOF SYSTEM - Techniques for improving the integrity and performance of biometric security processes on data processing devices are provided. An example of a method of determining a liveness of a biometric input according to the disclosure includes obtaining inquiry image information, obtaining enrollment image information, determining alignment information based on the inquiry image information and the enrollment image information, determining an overlap area based on the alignment information, determining anti-spoofing features based on the overlap area within the inquiry image information and the enrollment image information, and outputting a liveness score based on the anti-spoofing features. | 2021-02-04 |
| 20210034896 | LIVING BODY DETECTION METHOD, APPARATUS AND DEVICE - Methods, systems, and apparatus, including computer programs encoded on computer storage media, for liveness detection are provided. One of methods includes: displaying an image to a user, and capturing a face image of the user while displaying the image to the user; determining an eye image of an eye of the user based on the face image; extracting a to-be-verified image from the eye image of the user, wherein the to-be-verified image is reflection of the displayed image in the eye of the user; comparing the displayed image with the to-be-verified image to determine whether the to-be-verified image matches the displayed image; and performing liveness detection on the user based on a result of comparison. | 2021-02-04 |
| 20210034897 | INPUT/OUTPUT DEVICE, SCREEN CONTROL DEVICE, AND SCREEN CONTROL METHOD - An input/output device is configured to input or output operation information of an operator with a terminal device. The input/output device includes a display having a first display screen and a second display screen, an imaging device which is interposed between the first display screen and the second display screen and configured to capture an image of the operator positioned in the foreside of the terminal device, and a screen controller configured to control coordination of the first display screen and the second display screen and to thereby guide operator's eyes toward the imaging device interposed between the first display screen and the second display screen. | 2021-02-04 |
| 20210034898 | ENTERING OF HUMAN FACE INFORMATION INTO DATABASE - A processor chip circuit is provided, which is used for entering human face information into a database and includes a circuit unit configured to perform the steps of: videoing one or more videoed persons and extracting human face information of the one or more videoed persons from one or more video frames during the videoing; recording a voice of at least one of the one or more videoed persons during the videoing; performing semantic analysis on the recorded voice so as to extract respective information therefrom; and associating the extracted information with the human face information of the videoed person who has spoken the extracted information, and entering the associated information into the database. | 2021-02-04 |
| 20210034899 | Line-Based Feature Generation For Vision-Based Driver Assistance Systems And Methods - Embodiments provide line-based feature generation for vision-based driver assistance systems and methods. For one embodiment, a feature generator includes a circular buffer and a processor coupled to an image sensor. The circular buffer receives image data from the image sensor and stores N lines at a time of an image frame captured by the image sensor. The N lines of the image frame are less than all of the lines for the image frame. The processor receives the N lines from the circular buffer and stores one or more features generated from the N lines in a memory. Iterative blocks of N lines of image data are processed to complete processing of the full image frame, and multiple frames can be processed. The generated features are analyzed by a vision processor to identify, classify, and track objects for vision-based driver assistance and related vision-based assistance actions. | 2021-02-04 |
| 20210034900 | METHOD AND APPARATUS FOR EXTRACTING IMAGE DATA IN PARALLEL FROM MULTIPLE CONVOLUTION WINDOWS, DEVICE, AND COMPUTER-READABLE STORAGE MEDIUM - Embodiments of the present disclosure provide a method and apparatus for extracting image data in parallel from multiple convolution windows, a device, and a computer-readable storage medium. The method includes: dividing an image into multiple groups of convolution windows, where the multiple groups of convolution windows include a first group of convolution windows and a second group of convolution windows, and each group of convolution windows include multiple convolution windows. The method further includes extracting image data in parallel from multiple convolution windows in the first group of convolution windows by using multiple data processing units, and extracting, after the extraction of image data from the first group of convolution windows is completed, image data from multiple convolution windows in the second group of convolution windows in parallel by using the multiple data processing units. | 2021-02-04 |
| 20210034901 | TARGET OBJECT RECOGNITION METHOD AND APPARATUS, STORAGE MEDIUM, AND ELECTRONIC DEVICE - This application discloses a target object recognition method and apparatus, a storage medium, and an electronic device. The method includes: obtaining a first image and a second image, the first image being an image obtained by shooting a target scene under visible light, and the second image being an image obtained by shooting the target scene under infrared; determining a predicted infrared intensity value corresponding to a pixel point in the first image through a prediction model; obtaining a difference between an actual infrared intensity value of a pixel point in the second image and the infrared intensity value corresponding to the pixel point at the same location in the first image; and determining a pixel point in the second image of which the difference is greater than a first threshold as a pixel point at which the target object in the target scene is located. | 2021-02-04 |
| 20210034902 | TRAILER HITCHING ASSIST SYSTEM WITH TRAILER COUPLER DETECTION - A trailering assist system for a vehicle includes a camera disposed at a rear portion of a vehicle and having a field of view rearward of the vehicle, the field of view encompassing at least a portion of a trailer rearward of the vehicle. The camera captures image data, which is representative of at least a front profile of the trailer, which includes a trailer coupler of the trailer. An ECU includes an image processor that processes image data captured by the camera. The ECU, responsive to image processing of captured image data, determines a location of the front profile of the trailer relative to the vehicle. Responsive to determining the location of the front profile, the ECU determines a plurality of outline landmarks that represent a shape of the front profile of the trailer. Based on the determined outline landmarks, the ECU determines a location of the trailer coupler. | 2021-02-04 |
| 20210034903 | TRAILER HITCHING ASSIST SYSTEM WITH TRAILER COUPLER DETECTION - A trailer assist system for a vehicle includes a camera disposed at a rear portion of a vehicle and having a field of view exterior and at least rearward of the vehicle, the field of view encompassing at least a portion of a trailer coupler of a trailer stationary a distance from the vehicle. The camera captures image data that is representative of at least the trailer coupler of the trailer. An ECU includes an image processor operable to process image data captured by the camera. The ECU, responsive to image processing at the ECU of image data captured by the camera, determines a location of the trailer coupler using a detector model, which is based on an ensemble regression tree algorithm. | 2021-02-04 |
| 20210034904 | METHOD OF DETECTING REGION OF INTEREST ON BASIS OF GAZE DIRECTION AND ELECTRONIC DEVICE THEREFOR - Disclosed is an electronic device comprising a communication circuit, a memory, and a processor. The processor may be set to obtain an omnidirectional image, identify at least one face in the omnidirectional image, obtain at least one gaze direction from the at least one face, and generate at least one region of interest of the omnidirectional image on the basis of the at least one gaze direction. Also, various other embodiments specified throughout the present specification may also be possible. | 2021-02-04 |
| 20210034905 | APPARATUS, METHOD AND COMPUTER PROGRAM FOR ANALYZING IMAGE - The present disclosure relates to an image analysis method, system, and computer program. The image analysis method of the present disclosure includes: receiving a query image; extracting one or more regions of interest from the query image; calculating a first feature for each of the regions of interest by respectively applying the regions of interest to one or more ROI (region of interest) feature extraction models independently learned in order to extract features of the regions of interest; and calculating analysis values of the query image by applying the first features of the regions of interest to a pre-learned integration analysis model. According to the present disclosure, it is possible to reduce the influence on an analysis model by an error that training data created for map learning of an entire image may have, and it is also possible to increase learning accuracy and objectivity of a deep neural network. | 2021-02-04 |
| 20210034906 | DETERMINING RELATIVE REGIONS OF INTEREST IN IMAGES USING OBJECT DETECTION - In various examples, frames of a video may include a first visual object that may appear relative to a second visual object within a region of the frames. Once a relationship between the first visual object and the region is known, one or more operations may be performed on the relative region. For example, optical character recognition may be performed on the relative region where the relative region is known to contain textual information. As a result, the identification of the first visual object may serve as an anchor for determining the location of the relative region including the second visual object—thereby increasing accuracy and efficiency of the system while reducing run-time. | 2021-02-04 |
| 20210034907 | SYSTEM AND METHOD FOR TEXTUAL ANALYSIS OF IMAGES - Segmentation first breaks the images into segments or regions, with the segments of the region having text or symbols. The segmented image is separately applied to two different CNN-based models. Each model produces text boxes where potential text might exist. Then, a selective NMS algorithm is applied to the output of each model to produce a final group of text regions. These text regions are analyzed and actions taken. | 2021-02-04 |
| 20210034908 | INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING SYSTEM, AND CONTROL METHOD - An information processing system ( | 2021-02-04 |
| 20210034909 | SPATIAL TRANSFORMER MODULES - Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using an image processing neural network system that includes a spatial transformer module. One of the methods includes receiving an input feature map derived from the one or more input images, and applying a spatial transformation to the input feature map to generate a transformed feature map, comprising: processing the input feature map to generate spatial transformation parameters for the spatial transformation, and sampling from the input feature map in accordance with the spatial transformation parameters to generate the transformed feature map. | 2021-02-04 |
| 20210034910 | Adaptive Image Processing - An imaging device includes one or more processors; and a computer readable medium storing instructions that, when executed by the one or more processors, cause the imaging device to perform functions including: capturing a first image and thereafter a second image; making a determination of whether or not a difference between the first image and the second image is greater than a threshold value; generating a third image by processing the second image using an image processing algorithm that corresponds to the determination; and displaying the third image. | 2021-02-04 |
| 20210034911 | ITERATIVE MULTI-DIRECTIONAL IMAGE SEARCH SUPPORTING LARGE TEMPLATE MATCHING - Systems and methods which provide iterative multi-directional image searching are described. Embodiments utilize a multi-directional searching pattern for defining one or more searching areas within a source image. A searching area of embodiments provides a searching area within a source image for searching for a template image in the multiple directions of the multi-directional searching pattern. A location of the searching area within the source image may be updated iteratively, such as based upon motion vectors derived from the searching, until a template image match position is identified within the source image. Embodiments transform the template image and the searching area within a source image to 1D representations corresponding to each direction of the multi-directional image search. Embodiments accommodate rotation and scale variance of the subject (e.g., object of interest) of the template image within the source image. | 2021-02-04 |
| 20210034912 | DIFFERENTIATION DEVICE, DIFFERENTIATION METHOD FOR DEPRESSION SYMPTOMS, DETERMINATION METHOD FOR LEVEL OF DEPRESSION SYMPTOMS, STRATIFICATION METHOD FOR DEPRESSION PATIENTS, DETERMINATION METHOD FOR EFFECTS OF TREATMENT OF DEPRESSION SYMPTOMS, AND BRAIN ACTIVITY TRAINING DEVICE - Objective discrimination of a disease label of a depressive symptom with respect to an active state of a brain is achieved. One means for solving the problems of the present invention is to provide a discriminating device for assisting in determination of whether a subject has a depressive symptom. The discriminating device includes a storage device for storing information for identifying a classifier generated by classifier generation processing based on a signal obtained by using a brain activity detecting apparatus to measure, in advance and time-sequentially, a signal indicating a brain activity of a plurality of predetermined regions of each brain of a plurality of participants in a resting state, the plurality of participants including healthy individuals and patients with depression. The classifier is generated so as to discriminate a disease label of a depressive symptom based on a weighted sum of a plurality of functional connectivities selected by feature selection as being relevant to the disease label of the depressive symptom through machine learning from among functional connectivities of the plurality of predetermined regions. The discriminating device further includes a processor configured to execute discriminating processing of generating a classification result for the depressive symptom of the subject by using the classifier. | 2021-02-04 |
| 20210034913 | METHOD AND DEVICE FOR IMAGE PROCESSING, AND COMPUTER STORAGE MEDIUM - A method and device for image recognition and a storage medium are provided. The method includes: a target image is acquired; feature extraction processing is performed on the target image through convolutional layers in a neural network model to obtain feature maps, and Instance Normalization (IN) and Batch Normalization (BN) processing is performed on the feature maps to obtain a recognition result of the target image; and the recognition result of the target image is output. | 2021-02-04 |
| 20210034914 | DETECTION OF EMERGENCY VEHICLES - Aspects of the disclosure relate to detecting an emergency vehicle. For instance, a plurality of images may be taken from a perspective of an autonomous vehicle. One or more gates representing a region of interest at a respective distance from the vehicle may be generated for the images. A plurality of lights may be detected within the one or more gates. A first candidate emergency vehicle may be identified from a detected plurality of lights in one or more gates of one of the images, and a second candidate emergency vehicle may be identified from a detected plurality of lights in one or more gates of another of the images. The first and second candidate emergency vehicles are determined to be the same emergency vehicle and to be active. An operational system of the autonomous vehicle is controlled based on the determination that the given emergency vehicle is active. | 2021-02-04 |
| 20210034915 | METHOD AND APPARATUS FOR OBJECT RE-IDENTIFICATION - An object re-identification method performed by an object re-identification apparatus. The method includes detecting an object in a plurality of images; inferring object information including an attribute for the detected object; selecting an object having a same attribute as an identification target object from the inferred object information as a comparison target object; inferring a photographing angle of the selected comparison target object; selecting an identification candidate object from the comparison target object according to whether the inferred photographing angle is included in a predetermined angle range corresponding to the identification target object; and identifying whether the selected identification candidate object is matched with the identification target object. | 2021-02-04 |
| 20210034916 | SENSOR INFORMATION INTEGRATION SYSTEM, SENSOR INFORMATION INTEGRATION METHOD, AND PROGRAM - A system includes a first analysis apparatus, a second analysis apparatus, and a sensor fusion apparatus. The individual analysis apparatuses analyze outputs of their respective sensors and generate preciseness information about preciseness of the positions of their respective objects. The sensor fusion apparatus calculates the distance between a determination target object and an association candidate object per directional axis that is determined from a position of a sensor corresponding to the determination target object. The sensor fusion apparatus normalizes the distance calculated per directional axis by using the preciseness information corresponding to the determination target object. The sensor fusion apparatus calculates the distance between the determination target object and the association candidate object by using the distance normalized per directional axis as a post-normalization distance. The sensor fusion apparatus determines whether the association candidate object matches the determination target object by using the post-normalization distance. | 2021-02-04 |
| 20210034917 | ANOMALY DETECTION AND CLUSTERING IN FINANCIAL DATA CHANNEL MIGRATION - A method and system determines whether or not a new data acquisition process is working for individual financial accounts of users of a data management system. The method and system trains an analysis model with a machine learning process. The trained analysis model then analyzes financial data obtained by both an old data acquisition process and the new data acquisition process. The trained analysis model identifies whether the new data acquisition process is working properly based on the analysis. | 2021-02-04 |
| 20210034918 | Method and apparatus for detecting signal features - A measurement apparatus comprising an acquisition memory adapted to store data sections of at least one acquired measurement signal; a processor adapted to calculate a measurement parameter vector, v, for each data section of the acquired measurement signal; and a trained autoencoder neural network adapted to process the measurement parameter vectors, v, applied as input data to the trained autoencoder neural network to provide at a middle layer of said autoencoder neural network an encoded vector, h, with characteristic signal features of the acquired measurement signal. | 2021-02-04 |
| 20210034919 | METHOD AND APPARATUS FOR ESTABLISHING IMAGE SET FOR IMAGE RECOGNITION, NETWORK DEVICE, AND STORAGE MEDIUM - A method of establishing an image set for image recognition includes obtaining a single-label image set comprising an image annotated with a single label, and a multi-label image set comprising an image annotated with a plurality of labels; converting content of each label into a corresponding word identifier according to a semantic network, to obtain a word identifier set, a converted single-label image set, and a converted multi-label image set; and constructing a hierarchical semantic structure according to the word identifier set and the semantic network. The method also includes performing label supplementation on the image in the converted single-label image set to obtain a supplemented single-label image set; performing label supplementation on the supplemented single-label image set to obtain a final supplemented image set; and establishing a target multi-label image set to train an image recognition model by using the target multi-label image set. | 2021-02-04 |
| 20210034920 | ANNOTATION PIPELINE FOR MACHINE LEARNING ALGORITHM TRAINING AND OPTIMIZATION - Techniques are provided for enhancing the efficiency and accuracy of annotating data samples for supervised machine learning algorithms using an advanced annotation pipeline. According to an embodiment, a method can comprise collecting, by a system comprising a processor, unannotated data samples for input to a machine learning model and storing the unannotated data samples in an annotation queue. The method further comprises determining, by the system, annotation priority levels for respective unannotated data samples of the unannotated data samples, selecting, by the system from amongst different annotation techniques, one or more of the different annotation techniques for annotating the respective unannotated data samples based the annotation priority levels associated with the respective unannotated data samples. | 2021-02-04 |
| 20210034921 | SYSTEMS AND METHODS FOR GENERATING AUGMENTED TRAINING DATA FOR MACHINE LEARNING MODELS - Methods and systems are provided for augmenting ultrasound image training data, which may be used to train one or more machine learning models. One example method for augmenting ultrasound training data comprises, selecting an ultrasound image and a ground truth output corresponding to the ultrasound image, determining a first modification to apply to the ultrasound image, applying the first modification to the ultrasound image to produce an augmented ultrasound image, modifying the ground truth output based on the first modification to produce an augmented ground truth output corresponding to the augmented ultrasound image, and training a machine learning model using the augmented ultrasound image and the augmented ground truth output. In this way, a machine learning model may learn a more robust mapping from ultrasound image features to expected output, with less probability of overfitting, and with increased generalizability to noisy ultrasound images, or ultrasound images containing artifacts. | 2021-02-04 |
| 20210034922 | METHOD, ELECTRONIC DEVICE AND COMPUTER PROGRAM PRODUCT FOR PROCESSING DATA - A method comprises: generating, at a first computing device, a first set of gradient values associated with a data block processed by nodes of a machine learning model, the first set of gradient values being in a first data format; determining a first shared factor from the first set of gradient values, the first shared factor being in a second data format of a lower a precision than that of the first data format; and scaling the first set of gradient values with the first shared factor, to obtain a second set of gradient values having the second data format. In addition, the method comprises sending the second set of gradient values and the first shared factor to a second computing device; and, in response to receiving a third set of gradient values and a second shared factor from the second computing device, adjusting parameters of the machine learning model. | 2021-02-04 |
| 20210034923 | SYSTEM, METHOD, AND NON-TRANSITORY STORAGE MEDIUM - A system manages a prediction model generated with input data including a plurality of parameters in a learning process and includes: a first calculation unit configured to calculate a learning contribution ratio of each of the plurality of parameters from a learning result; a first specifying unit configured to specify parameters of which the learning contribution ratios are low from the learning contribution ratios calculated by the first calculation unit; a first instruction unit configured to give a relearning instruction using input data excluding the parameters specified by the first specifying unit; a second specifying unit configured to specify configurations corresponding to the parameters specified by the first specifying unit; and a first issuing unit configured to issue a stop command for the configuration specified by the second specifying unit. | 2021-02-04 |
| 20210034924 | SYSTEMS AND METHODS FOR IMPLEMENTING AN INTELLIGENT TUNING OF MULTIPLE HYPERPARAMETER CRITERIA OF A MODEL CONSTRAINED WITH METRIC THRESHOLDS - Systems and methods for tuning hyperparameters of a model include receiving a tuning request for tuning hyperparameters, the tuning request includes a first and a second objective function for the machine learning model. The first and second objective functions may output metric values that do not improve uniformly. Systems and methods additionally include defining a joint tuning function that is based on a combination of the first and second objective functions; executing a tuning operation; identifying a Pareto efficient frontier curve defined by a plurality of distinct hyperparameter values; applying metric thresholds to the Pareto efficient frontier curve; demarcating the Pareto efficient frontier curve into at least a first infeasible section and a second feasible section; searching the second feasible section of the Pareto efficient frontier curve for one or more proposed hyperparameter values; and identifying at least a first set of proposed hyperparameter values based on the search. | 2021-02-04 |
| 20210034925 | OBSTACLE RECOGNITION ASSISTANCE DEVICE, OBSTACLE RECOGNITION ASSISTANCE METHOD, AND STORAGE MEDIUM - An obstacle recognition assistance device includes: an acquisition unit that acquires a first recognition result and sensing information, the first recognition result being a result based on a first recognition process performed using the sensing information acquired by an obstacle sensing device, the first recognition result indicating whether or not an obstacle is present, the first recognition result being determined to be incorrect; and a recording unit that verifies whether the first recognition result is incorrect based on the first recognition result and a second recognition result, the second recognition result being a result based on a second recognition process performed using the sensing information, the second recognition process indicating whether or not the obstacle is present. | 2021-02-04 |
| 20210034926 | FEEDBACK DRIVEN DECISION SUPPORT IN PARTIALLY OBSERVABLE SETTINGS - A novel formulation called the Context-Attentive Bandit with Observations (CABO) is described, where only a limited number of features can be accessed by the learner. The present invention is applicable to many problems including problems arising in clinical settings and dialog systems where it is not possible to reveal the whole feature set. The present invention adapts the standard contextual bandit algorithm known as Thompson Sampling with a novel algorithm, we call Context-Attentive Thompson Sampling with Observations (CATSO). Experimental results are included to demonstrate its effectiveness including a regret analysis and an empirical evaluation demonstrating advantages of the disclosed novel approach on several real-life datasets. | 2021-02-04 |
| 20210034927 | APPARATUS AND METHOD FOR PERSON DETECTION, TRACKING, AND IDENTIFICATION UTILIZING WIRELESS SIGNALS AND IMAGES - An apparatus comprising a wireless transceiver configured to communicate packet data with a mobile device associated with one or more persons in a vicinity of the wireless transceiver, and a controller in communication with the wireless transceiver and a camera, the controller configured to receive a plurality of packet data from one or more person, wherein the packet data includes at least amplitude information associated with the wireless channel communicating with the wireless transceiver and receive images from the camera containing trajectory of motion of individuals, performs detection, tracking, and pseudo-identification of individuals by fusing motion trajectories from wireless signals and camera images. | 2021-02-04 |
| 20210034928 | COMBINATORIAL BAYESIAN OPTIMIZATION USING A GRAPH CARTESIAN PRODUCT - Certain aspects provide a method for determining a solution to a combinatorial optimization problem, including: determining a plurality of subgraphs, wherein each subgraph of the plurality of subgraphs corresponds to a combinatorial variable of the plurality of combinatorial variables; determining a combinatorial graph based on the plurality of subgraphs; determining evaluation data comprising a set of vertices in the combinatorial graph and evaluations on the set of vertices; fitting a Gaussian process to the evaluation data; determining an acquisition function for vertices in the combinatorial graph using a predictive mean and a predictive variance from the fitted Gaussian process; optimizing the acquisition function on the combinatorial graph to determine a next vertex to evaluate; evaluating the next vertex; updating the evaluation data with a tuple of the next vertex and its evaluation; and determining a solution to the problem, wherein the solution comprises a vertex of the combinatorial graph. | 2021-02-04 |
| 20210034929 | INTER-CLASS ADAPTIVE THRESHOLD STRUCTURE FOR OBJECT DETECTION - Provided herein are systems and methods for applying adaptive classes thresholds to enhance object detection Machine Learning (ML) models by receiving a plurality of labeled feature vectors extracted from a plurality of images associated with a plurality of objects, one or more subsets of the plurality of feature vectors are associated with respective object(s) and labeled accordingly, computing an adaptive threshold for each object in a plurality of iterations, each iteration comprising: (1) computing deviation of a respective feature vector of the subset from an aggregated feature vector, (2) computing, in case the deviation is within a predefined value, a threshold enclosing the respective feature vector, and (3) adjusting the adaptive threshold to enclose the threshold of the respective feature vector and outputting the adaptive threshold(s) for classifying unlabeled feature vectors to class(s) of respective object(s) associated with the adaptive threshold(s) in which the unlabeled feature vectors fall. | 2021-02-04 |
| 20210034930 | COLLATION/RETRIEVAL SYSTEM, COLLATION/RETRIEVAL SERVER, IMAGE FEATURE EXTRACTION APPARATUS, COLLATION/RETRIEVAL METHOD, AND PROGRAM - The present invention is a collation/retrieval system collating a product manufactured by or delivered from a producer or a distributor with a product to be collated comprising: a storage unit that stores an image feature of a predetermined collation area of the product determined in advance at a position relative to a reference section common in every product; a to-be-collated product feature extraction unit that receives an image of the product to be collated and detecting the reference section of the product from the received image to extract an image feature of the collation area determined by reference to the reference section; and a collation unit that collates the stored image feature with the image feature of the collation area of the product to be collated. | 2021-02-04 |
| 20210034931 | IMAGE FORMING APPARATUS - An image forming apparatus includes a plurality of paper feed portions, an image forming portion, a storage portion that stores, for each of the paper feed portions, sheet delay information indicating an occurrence frequency of a sheet delay that is a phenomenon in which a time required for arrival of a front end of the sheet at a prescribed position from the start of feeding of the sheet is larger than a threshold time, and a control portion that, based on the sheet delay information on the each of the paper feed portions, recognizes the occurrence frequency of a sheet delay of each of the plurality of paper feed portions and preferentially sets, as a paper feed source, one of the paper feed portions whose value of the occurrence frequency of a sheet delay is equal to or less than a prescribed value. | 2021-02-04 |
| 20210034932 | TRANSACTION CARD WITH INTEGRATED STAND - The disclosed embodiments generally relate to a transaction card with an integrated card stand. The transaction card comprises a substrate having a card thickness, and first and second card surfaces, the second card surface being substantially coextensive with the first card surface. The transaction may further comprise a card stand pivotable relative to the second card surface and having a stand thickness. | 2021-02-04 |
| 20210034933 | TWO DIMENSIONAL BARCODE PROVIDED WITH SURROUNDING DYNAMIC ENVIRONMENTAL INDICATOR AND COLOR CALIBRATION REFERENCE - An indicator includes a substrate, a two-dimensional barcode symbol provided on the substrate, a space on the substrate surrounding the barcode symbol, and a historical environmental condition indicator on the substrate. The two-dimensional barcode symbol may be printed on the substrate, preferably in a rectangular, and more preferably in a square configuration. The two-dimensional barcode is printed preferably in black and white and more preferably according to a certain standard. The space on the substrate surrounding the barcode symbol has four quadrants. Additionally, the historical environmental condition indicator is preferably a historical temperature indicator, which occupies space on the substrate in at least two of the four quadrants. The historical environmental condition indicator preferably occupies three and more preferably occupies four of the four quadrants. | 2021-02-04 |
| 20210034934 | SYSTEM AND METHOD FOR TRASMITTING FINANCIAL INFORMATION VIA COLOR MATRIX CODE - An apparatus may include a memory to store instructions; and processing circuitry, coupled with the memory, operable to execute the instructions. When executed, the instructions may cause the processing circuitry to identify a matrix code; read a first layer of the matrix code, the first layer comprising a first account identifier associated with an account, wherein the first layer corresponds to a first color channel; and read a second layer of the matrix code, the second layer comprising a set of account data, associated with the account, wherein the second layer corresponds to a second color channel, different than the first color channel. | 2021-02-04 |
| 20210034935 | COMPOSITE SUBSTRATES FOR SAW TAGS OR RFID AND SENSORS APPLICATIONS - A surface acoustic wave tag device is disclosed, comprising: an acoustic wave propagating substrate, at least one transducer structure comprising inter-digitated comb electrodes, and at least one reflecting means, the reflecting means comprising at least one reflector, wherein the acoustic wave propagation substrate is a composite substrate comprising a base substrate and a piezoelectric layer, wherein the crystallographic orientation of the piezoelectric layer with respect to the base substrate is such that the propagation of a shear wave inside the piezoelectric layer and in the direction of propagation corresponding to the acoustic wave is enabled. A physical quantity determining device and a fabrication method of such surface acoustic wave tag device are also disclosed. | 2021-02-04 |
| 20210034936 | DEMODULATOR AND METHOD OF DEMODULATING ASK SIGNAL - A new demodulator with consistent sensibility to signals received from different directions, low power consumption, and low manufacturing cost is provided. The demodulator may include a first demodulator branch and a second demodulator branch electrically connected in parallel, and a DC circuit to provide DC power to the demodulator. The DC circuit has a first diode and a second diode electrically connected in series between a DC power supply Vcc and the ground. The second demodulator branch can share a low pass filter and a DC blocking capacitor of the first demodulator branch for example, and can multiplex or reuse a bias current from the first demodulator branch. | 2021-02-04 |
| 20210034937 | WIRELESS DEVICE CONFIGURED FOR POWERING TRANSMISSIONS WITH HARVESTED ENERGY - Embodiments of the present disclosure generally relate to a wireless identification tag configured to harvest ambient energy and transmit an identification signal intermittently, and system and methods for use thereof. In one implementation, the tag may include an antenna configured to receive ambient energy. The tag may also include an energy storage component configured to aggregate and store the received ambient energy. The tag may also include a transmitter electrically connected to the energy storage component and configured to transmit the identification signal. The tag may also include a circuit connected to the transmitter and configured to implement an identification transmission rule, to cause the transmitter to delay sending the identification signal even when sufficient energy for transmission of the identification signal is aggregated and stored in the energy storage component. | 2021-02-04 |
| 20210034938 | MANAGEMENT SYSTEM FOR GAME TOKEN COIN - In a management system of a game token coin, one RFID tag having a data non-rewritable region and a data rewritable region is imparted to the game token coin. The location relating information of the game token coin is stored in the data rewritable region. The system includes a writer configured to additionally write the location relating information, a reader configured to read the location relating information from a plurality of the game token coins, and a management controller configure to examine the information obtained from the reader. The management controller includes a function that determines whether a history of the location relating information obtained from the reader has no abnormality. | 2021-02-04 |
| 20210034939 | RFIC MODULE AND RFID TAG - In an impedance matching circuit, first to fourth inductors are formed by a conductor pattern of a coil shape, and a fifth inductor is formed by a conductor pattern of an unwound shape. The first inductor and the third inductor are respectively formed in different layers of a substrate and are arranged in such a relationship as to have coil openings overlapping each other. Moreover, the second inductor and the fourth inductor are respectively formed in different layers of the substrate and are arranged in such a relationship as to have coil openings overlapping each other. Two coils interposing the mounting position of the RFIC are in a 180° rotational symmetry relationship. | 2021-02-04 |
| 20210034940 | Cards & Documents - One 16 digits smart card shaped like a regular bank card in length and width with chip and swiping abilities which is divided into several different separate sectors. e.g. bank cards, credit cards, cash transactions, medical insurance usage and other personal documentation of information that are all stored on one smart card. This smart card is configured to hold financial information, general information, and documentation of a variety of various categories of different documents on its database. All parts can be accessible by one four-digit secret security code to ensure the safety of the card's information. | 2021-02-04 |
| 20210034941 | METHOD OF MANUFACTURING RFID ACCESS CARD FROM NOVEL MATERIALS - A method, apparatus, and product-by-process, for preparing an RFID access card constructed out of wood is disclosed. The method involved preparing two wooden surfaces of appropriate thickness, preparing an RFID chip emitting an appropriate frequency, and pressing the wooden surfaces and the RFID chip together through the use of adhesive(s) and a hot press and a cold press. The apparatus included an RFID access card constructed out of at least two wooden surfaces where an RFID chip is disposed between the two surfaces. The two wooden surfaces are attached with a hot melt adhesive. | 2021-02-04 |
| 20210034942 | SELF-CENTERED INLAY AND CORE LAYER FOR INFORMATION CARRYING CARD, PROCESS AND RESULTING PRODUCTS - The disclosure provides a method for forming a core layer for at least one information carrying card, and resulting products. The method includes forming an inlay layout, and dispensing a crosslinkable polymer composition over the inlay layout and contacting the inlay layer so as to form the core layer of the information carrying card. The inlay layout includes at least one inlay layer coupled with a first thermoplastic layer. The first thermoplastic layer comprises a thermoplastic material, and defines at least one hole therein. The at least one inlay layer is disposed at least partially inside a respective hole. | 2021-02-04 |
| 20210034943 | NON-CONTACT COMMUNICATION MEDIUM AND RECORDING MEDIUM CARTRIDGE - A non-contact communication medium according to an embodiment of the present technology is a non-contact communication medium for a recording medium cartridge, including: a circuit component; a support substrate; and an antenna coil. The circuit component has a memory unit capable of storing management information relating to the recording medium cartridge therein. The support substrate supports the circuit component. The antenna coil includes a coil unit that is electrically connected to the circuit component and formed on the support substrate, an inductance value of the coil unit being 0.3 μH or more and 2.0 μH or less. | 2021-02-04 |
| 20210034944 | SYSTEM AND METHOD FOR PROPAGATING INFORMATION USING MODIFIED NUCLEIC ACIDS - A method is provided for improving a nucleic acid-based molecular computing system comprised of (i) a nucleic acid structure, (ii) at least one polynucleotide displacement molecule that can bind with the nucleic acid structure under hybridizing conditions, and (iii) a clashing polynucleotide molecule that competes with the polynucleotide displacement molecule for binding the nucleic acid structure under the hybridizing conditions The method entails incorporation of chemical modification that inhibits the binding of the clashing molecule and the nucleic acid structure or facilitates the binding of the displacement molecule and the nucleic structure. | 2021-02-04 |
| 20210034945 | PERSONALIZED COMPLIMENTARY ITEM RECOMMENDATIONS USING SEQUENTIAL AND TRIPLET NEURAL ARCHITECTURE - A system and method of generating complimentary items from a catalog of items is disclosed. A plurality of item attributes for each of a plurality of items is received and a multimodal embedding representative of the plurality of attributes is generated for each of the plurality of items. The multimodal embedding is configured to predict at least a subset of the received plurality of item attributes for each of the plurality of items. A triplet network including a node representative of each of the plurality of items is generated. The triplet network is generated based on the multimodal embedding for each of the plurality of items. A plurality of complimentary items is generated from the plurality of items. The plurality of complimentary items are selected by the triplet network based on an anchor item selection received from a user. | 2021-02-04 |
| 20210034946 | RECOGNIZING PROBLEMS IN PRODUCTIVITY FLOW FOR PRODUCTIVITY APPLICATIONS - Recognizing problems in productivity flow for productivity applications can be accomplished by receiving a user profile and actions for a timeframe, where each action comprises an activity and timing; and identifying a grammar that captures action semantics and sequences in which they are executed that is similar to at least the actions for the timeframe, where the grammar corresponds to a help case. When the grammar is identified, corresponding help content for the help case can be retrieved and provided to the user. | 2021-02-04 |
| 20210034947 | BAYESIAN NEURAL NETWORK AND METHODS AND APPARATUS TO OPERATE THE SAME - Methods, apparatus, systems, and articles of manufacture providing an improved Bayesian neural network and methods and apparatus to operate the same are disclosed. An example apparatus includes an oscillator to generate a first clock signal; a resistive element to adjust a slope of a rising edge of a second clock signal; a voltage sampler to generate a sample based on at least one of (a) a first voltage of the first clock signal when a second voltage of the second clock signal satisfies a threshold or (b) a third voltage of the second clock signal when a fourth voltage of the first clock signal satisfies the threshold; and a charge pump to adjust a weight based on the sample, the weight to adjust data in a model. | 2021-02-04 |
| 20210034948 | METADATA-BASED SCIENTIFIC DATA CHARACTERIZATION DRIVEN BY A KNOWLEDGE DATABASE AT SCALE - A metadata-based scientific data characterization method, system, and computer program product include requesting a user input for a task to specify a rule for the task to determine a quality and a relationship of a data file in a data file database based on metadata associated with the data file, processing a user feedback of results using the rule run on the data file database and tracking the user feedback on the results in order to learn from the user feedback, and based on the learning, creating a modified rule to determine a quality and a relationship of a second data | 2021-02-04 |
| 20210034949 | SYSTEMS AND METHODS FOR PREDICTING INFORMATION HANDLING RESOURCE FAILURES USING DEEP RECURRENT NEURAL NETWORK WITH A MODIFIED GATED RECURRENT UNIT HAVING MISSING DATA IMPUTATION - A method may include receiving telemetry data associated with one or more information handling resources, receiving failure statistics associated with the one or more information handling resources, merging the telemetry data and the failure statistics to create training data, and implementing a gated recurrent unit to: (i) impute missing values from the training data and (ii) train a pattern recognition engine configured to predict a failure status of an information handling resource from operational data associated with the information handling resource. | 2021-02-04 |
| 20210034950 | METHOD FOR IMPLEMENTING NEURAL NETWORK MODEL IN HETEROGENEOUS COMPUTING PLATFORM AND APPARATUS FOR PERFORMING THE SAME - A method and apparatus for implementing a neural network model in a heterogeneous computing platform are disclosed. The method includes partitioning a neural network model into first sub-models based on a partition standard, obtaining second sub-models by merging at least a portion of the first sub-models based on characteristics of the first sub-models, and deploying the second sub-models. | 2021-02-04 |
| 20210034951 | MODULAR DISTRIBUTED ARTIFICIAL NEURAL NETWORKS - A modular neural network system comprising: a plurality of neural network modules; and a controller configured to select a combination of at least one of the neural network modules to construct a neural network, dedicated for a specific task. | 2021-02-04 |
| 20210034952 | PERTURBATIVE NEURAL NETWORK - Perturbative neural networks are comprised of one or more modules, called perturbation layers which can be used as an alternative to a convolutional layer. The perturbation layer does away with convolution in the traditional sense and instead computes its response as a weighted linear combination of non-linearly activated additive noise perturbed inputs. | 2021-02-04 |
| 20210034953 | RECONFIGURABLE FINFET-BASED ARTIFICIAL NEURON AND SYNAPSE DEVICES - A semiconductor device that implements artificial neurons and synapses together on the semiconductor device includes a plurality of fins formed on the semiconductor device, and a plurality of gates formed around the plurality of fins to form a plurality of fin field-effect transistors (FinFETs). The plurality of FinFETs may form one or more artificial synapses and one or more artificial neurons. Each of the one or more artificial synapses may include two or more of the plurality of gates. Each of the one or more artificial neurons comprises one of the plurality of gates. | 2021-02-04 |
| 20210034954 | MAGNETIC EFFECT ARTIFICIAL INTELLIGENCE SYSTEM - A magnetic effect artificial intelligence system composed of at least subsystems of tightly adherent hexagon and stereoscopic magnetic effect artificial neurons, input pre-processing unit, output unit, circuits, interconnections, electronics, electromagnetic components. | 2021-02-04 |
| 20210034955 | QUANTIZING NEURAL NETWORKS USING APPROXIMATE QUANTIZATION FUNCTION - Some embodiments of the invention provide a novel method for training a quantized machine-trained network. Some embodiments provide a method of scaling a feature map of a pre-trained floating-point neural network in order to match the range of output values provided by quantized activations in a quantized neural network. A quantization function is modified, in some embodiments, to be differentiable to fix the mismatch between the loss function computed in forward propagation and the loss gradient used in backward propagation. Variational information bottleneck, in some embodiments, is incorporated to train the network to be insensitive to multiplicative noise applied to each channel. In some embodiments, channels that finish training with large noise, for example, exceeding 100%, are pruned. | 2021-02-04 |
| 20210034956 | MINIMUM MEMORY DIGITAL CONVOLVER - An advancement over previous techniques, input data frames are streamed row-by-row to a convolver that can calculate and stream out convolution values without storing the convolution values. For many convolution operations only a single row of partial sums is stored. As input data values are received, they can be multiplied by kernel values and accumulated as partial sums until a convolution value is calculated. Convolution values can be clocked out of the convolver as soon as they are produced, thereby freeing the memory cell for use in calculating a different convolution sum. Clocking out convolution values as soon as they become available produces an output data stream of convolution values. By freeing memory cells and then reusing them as soon as possible, a convolver with a small, perhaps minimum, number of memory cells amount of memory is realized. | 2021-02-04 |
| 20210034957 | NEURAL NETWORK ACCELERATION SYSTEM AND OPERATING METHOD THEREOF - Disclosed are a neural network acceleration system and an operating method of the same. The neural network acceleration system includes a first memory module that generates a first reduced embedding segment through a tensor operation, based on a first segment of a first embedding and a second segment of a second embedding, a second memory module that generates a second reduced embedding segment through the tensor operation, based on a third segment of the first embedding and a fourth segment of the second embedding, and a processor that processes a reduced embedding including the first reduced embedding segment and the second reduced embedding segment, based on a neural network algorithm. | 2021-02-04 |
| 20210034958 | CONFIGURABLE PROCESSOR FOR IMPLEMENTING CONVOLUTION NEURAL NETWORKS - Configurable processors for implementing CNNs are provided. One such configurable CNN processor includes a plurality of core compute circuitry elements, each configured to perform a CNN function in accordance with a preselected dataflow graph, an active memory buffer, a plurality of connections between the active memory buffer and the plurality of core compute circuitry elements, each established in accordance with the preselected dataflow graph, a plurality of connections between the plurality of core compute circuitry elements, each established in accordance with the preselected dataflow graph, wherein the active memory buffer is configured to move data between the plurality of core compute circuitry elements via the active memory buffer in accordance with the preselected dataflow graph. | 2021-02-04 |
| 20210034959 | CONTINUOUSLY LEARNING AND OPTIMIZING ARTIFICIAL INTELLIGENCE (AI) ADAPTIVE NEURAL NETWORK (ANN) COMPUTER MODELING METHODS AND SYSTEMS - Continuously learning and optimizing artificial intelligence (AI) adaptive neural network (ANN) computer modeling methods and systems, designated human affect computer modeling (HACM) or affective neuron (AN), and, more particularly, to AI methods, systems and devices that can recognize, interpret, process and simulate human reactions and affects such as emotional responses to internal and external sensory stimuli, that provides real-time reinforcement learning modeling that reproduces human affects and/or reactions, wherein the human affect modeling (HACM) can be used singularly or collectively to modeling and predict complex human reactions and affects. | 2021-02-04 |
| 20210034960 | INTELLIGENT RETRAINING OF DEEP LEARNING MODELS - In an approach to deriving highly accurate models, one or more computer processors train a set of machine learning models utilizing a training set and a deep learning algorithm; generate one or more feedback data sets for each model in the set of trained models; rank each model in the set of trained models based on the generated feedback data sets; dynamically adjust one or more thresholds, that initiate a retraining or deployment of one or more ranked models, based, at least in part, on one or more production environment requirements; responsive to exceeding one or more adjusted thresholds, automatically deploy one or more ranked models to one or more deployment environments based, at least in part, on the ranking of the one or more trained models; responsive to not exceeding one or more adjusted thresholds, retrain each model in the set of trained models. | 2021-02-04 |
| 20210034961 | DEFECT REMOVAL FROM MANUFACTURED OBJECTS HAVING MORPHED SURFACES - Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided repair of physical structures include: generating a two dimensional difference image from a first three dimensional model of at least one actual three dimensional surface of a manufactured object, and a second three dimensional model of at least one source three dimensional surface used as input to a manufacturing process that generated the manufactured object; obtaining from an image-to-image translation based machine learning algorithm, trained using pairs of input images representing deformed and deformed plus surface defected added versions of a nominal three dimensional surface, a translated version of the two dimensional image; generating from the translated version of the two dimensional image a third three dimensional model of at least one morphed three dimensional surface corresponding to the at least one source three dimensional surface. Further, defects can be removed based on the third three dimensional model. | 2021-02-04 |
| 20210034962 | LEARNING AND RECALL IN SPIKING NEURAL NETWORKS - Methods and apparatus are provided for memorizing data signals in a spiking neural network. For each data signal, such a method includes supplying metadata relating to the data signal to a machine learning model trained to generate an output signal, indicating a relevance class for a data signal, from input metadata for that data signal. The method includes iteratively supplying the data signal to a sub-assembly of neurons, interconnected via synaptic weights, of a spiking neural network and training the synaptic weights to memorize the data signal in the sub-assembly. The method further comprises assigning neurons of the network to the sub-assembly in dependence on the output signal of the model such that more relevant data signals are memorized by larger sub-assemblies. The data signal memorized by a sub-assembly can be subsequently recalled by activating neurons of that sub-assembly. | 2021-02-04 |
| 20210034963 | IDENTIFYING FRICTION POINTS IN CUSTOMER DATA - Aspects of the present disclosure relate to identifying friction points in customer data. In some embodiments, identifying friction points can include receiving a set of input sequence data and predicted class labels for the set of input sequence data; selecting input sequences, from the set of input sequence data, that have class labels matching a ground truth class label; reducing the selected sequences to anchor points; and grouping the reduced selected sequences into critical data set signatures using discriminatory subsequence mining. | 2021-02-04 |
| 20210034964 | ANNOTATING CUSTOMER DATA - Aspects of the present disclosure relate to annotating or tagging customer data. In some embodiments, the annotating can include summarizing touchpoints into k-hot encoding feature vectors, mapping the feature vectors onto an embedding layer, predicting a hierarchical data sequence using the embedding layer and the feature vectors, extracting the feature vectors that are most influential in predicting the embedding layer, and outputting the touchpoints associated with the most influential feature vectors. | 2021-02-04 |
| 20210034965 | OUT-OF-DOMAIN ENCODER TRAINING - A computer-implemented method includes using an embedding network to generate prototypical vectors. Each prototypical vector is based on a corresponding label associated with a first domain. The computer-implemented method also includes using the embedding network to generate an in-domain test vector based on at least one data sample from a particular label associated with the first domain and using the embedding network to generate an out-of-domain test vector based on at least one other data sample associated with a different domain. The computer-implemented method also includes comparing the prototypical vectors to the in-domain test vector to generate in-domain comparison values and comparing the prototypical vectors to the out-of-domain test vector to generate out-of-domain comparison values. The computer-implemented method also includes modifying, based on the in-domain comparison values and the out-of-domain comparison values, one or more parameters of the embedding network. | 2021-02-04 |
| 20210034966 | METHOD AND DEVICE FOR NATURAL LANGUAGE PROCESSING AND STORAGE MEDIUM - A method for natural language, includes: determining a slot tagging result output by a Bi-directional Long Short-Term Memory-Conditional Random Field algorithm (BiLSTM-CRF) model after slot tagging on conversation data input by a user; determining reward information based on the slot tagging result and a reward of the user for the slot tagging result; and performing reinforcement learning on the BiLSTM-CRF model according to the reward information. | 2021-02-04 |
| 20210034967 | ENVIRONMENT CONTROLLER AND METHODS FOR VALIDATING AN ESTIMATED NUMBER OF PERSONS PRESENT IN AN AREA - Methods and environment controller for validating an estimated number of persons present in an area. The controller determines a temperature measurement in the area, a carbon dioxide (CO2) level measurement in the area, a humidity level measurement in the area, and the estimated number of persons present in the area based on data generated by an occupancy sensor. The controller executes a neural network inference engine for generating outputs based on inputs, using a predictive model comprising weights of a neural network. The inputs include the temperature, CO2 level and humidity level measurements, and the estimated number of persons. The outputs include an inferred temperature, an inferred CO2 level, an inferred humidity level, and an inferred number of persons. The controller applies a validation algorithm to the inputs and outputs of the neural network inference engine, to determine if the estimated number of persons is accurate or not. | 2021-02-04 |
| 20210034968 | NEURAL NETWORK LEARNING APPARATUS FOR DEEP LEARNING AND METHOD THEREOF - Disclosed is a neural network learning apparatus for deep learning and a method thereof. A neural network learning apparatus for deep learning according to an embodiment of the present disclosure includes an input interface, a memory, and a learning processor for applying a Gradient Descent algorithm to a neural network model, and the learning processor may transform a cumulative change function of the gradient for an error function into an inverse square root function in the Gradient Descent algorithm, and operate an inverse square root approximate value by using a Newton-Raphson method for the transformed inverse square root function. The neural network learning apparatus for deep learning of the present disclosure may be connected or converged with an Artificial Intelligence module, an Unmanned Aerial Vehicle (UAV), a robot, an Augmented Reality (AR) apparatus, a Virtual Reality (VR), or a 5G network service-related apparatus, etc. | 2021-02-04 |
| 20210034969 | TRAINING AN UNSUPERVISED MEMORY-BASED PREDICTION SYSTEM TO LEARN COMPRESSED REPRESENTATIONS OF AN ENVIRONMENT - Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a memory-based prediction system configured to receive an input observation characterizing a state of an environment interacted with by an agent and to process the input observation and data read from a memory to update data stored in the memory and to generate a latent representation of the state of the environment. The method comprises: for each of a plurality of time steps: processing an observation for the time step and data read from the memory to: (i) update the data stored in the memory, and (ii) generate a latent representation of the current state of the environment as of the time step; and generating a predicted return that will be received by the agent as a result of interactions with the environment after the observation for the time step is received. | 2021-02-04 |
| 20210034970 | DISTRIBUTED TRAINING USING ACTOR-CRITIC REINFORCEMENT LEARNING WITH OFF-POLICY CORRECTION FACTORS - Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network used to select actions to be performed by an agent interacting with an environment. In one aspect, a system comprises a plurality of actor computing units and a plurality of learner computing units. The actor computing units generate experience tuple trajectories that are used by the learner computing units to update learner action selection neural network parameters using a reinforcement learning technique. The reinforcement learning technique may be an off-policy actor critic reinforcement learning technique. | 2021-02-04 |
| 20210034971 | METHOD AND SYSTEM WITH NEURAL NETWORK MODEL UPDATING - A method of updating a neural network model by a terminal device, includes training a local model using a local data set collected by a terminal device to generate a trained local model; receiving, from a server, an independent identically distributed (i.i.d.) global data set, the i.i.d. global data set being a data set sampled for each class in a plurality of predefined classes; implementing the trained local model by inputting the i.i.d. global data set and transmitting final inference results of the implemented trained local model to the server; and receiving, from the server, a global model updated based on the final inference results of the inference. | 2021-02-04 |
| 20210034972 | BATCH NORMALIZATION LAYER TRAINING METHOD - Disclosed is a batch normalization layer training method, which may be used in a neural network learning apparatus having limited operational processing capability and storage space. A batch normalization layer training method according to an embodiment of the present disclosure may perform batch normalization transform by setting the gradients of the standard deviation and the mean of the loss function to zero, and applying a normalized statistic value obtained from an initial neural network or a previous neural network to the gradient of the loss function. The neural network learning apparatus of the present disclosure may be connected or converged with an Artificial Intelligence module, an Unmanned Aerial Vehicle (UAV), a robot, an Augmented Reality (AR) apparatus, a Virtual Reality (VR), a 5G network service-related apparatus, etc. | 2021-02-04 |
| 20210034973 | TRAINING NEURAL NETWORKS USING LEARNED ADAPTIVE LEARNING RATES - Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes training the neural network for one or more training steps in accordance with a current learning rate; generating a training dynamics observation characterizing the training of the trainee neural network on the one or more training steps; providing the training dynamics observation as input to a controller neural network that is configured to process the training dynamics observation to generate a controller output that defines an updated learning rate; obtaining as output from the controller neural network the controller output that defines the updated learning rate; and setting the learning rate to the updated learning rate. | 2021-02-04 |
| 20210034974 | DEVICES AND METHODS FOR REINFORCEMENT LEARNING VISUALIZATION USING IMMERSIVE ENVIRONMENTS - Disclosed are systems, methods, and devices for generating a visualization of a deep reinforcement learning (DRL) process. State data is received, reflective of states of an environment explored by an DRL agent, each state corresponding to a time step. For each given state, saliency metrics are calculated by processing the state data, each metric measuring saliency of a feature at the time step corresponding to the given state. A graphical visualization is generated, having at least two dimensions in which: each feature of the environment is graphically represented along a first axis; and each time step is represented along a second axis; and a plurality of graphical markers representing corresponding saliency metrics, each graphical marker having a size commensurate with the magnitude of the particular saliency metric represented, and a location along the first and second axes corresponding to the feature and time step for the particular saliency metric. | 2021-02-04 |
