Abstract: A scanning system comprises an intraoral scanner to capture scan data of a dental site and a computing device to generate a 3D rendering of the dental site. The scanner comprises one or more input devices configured to provide manual interaction with the computing device, where: a first activation of the input device(s) causes the scanning system to enter an inspection mode in which a view of the 3D rendering is adjustable based on movements of the intraoral scanner; and a second activation of the one or more input devices causes the scanning system to enter an overlay mode, wherein the computing device is to present a menu visualized on top of the 3D rendering on the display while the scanning system is in the overlay mode, the menu comprising one or more instructions for use of the intraoral scanner to control a user interface of the scanning system.

Abstract: A method reduces storage usage during processing of a neural network performed by an information processing apparatus comprising a storage. The network may be represented by operators that operate on an input feature map and generate an output feature map. A representation of the network is generated as a linear sequence of operators. Operators are identified in the sequence that cannot form part of a cascade and are to be processed with the entire input and output feature map of the respective operator in the storage. The method forms one or more cascades of two or more successive operators in the sequence for which the input feature map of each operator of the cascade is processed in portions, which portions are less than the entire input feature map.

Abstract: Described herein are examples of interfaces for designing and/or configuring custom assisted defect recognition (ADR) systems and/or workflows. In some examples, the ADR systems and/or workflows may be designed and/or configured using visual tools, such as, for example, node based processing tools. In some examples, the ADR workflows may be used to analyze two dimensional (2D) and/or three dimensional (3D) image scans, such as might be generated by industrial X-ray scanning systems.

Abstract: An integrated circuit includes an analog network of analog components fabricated by a method. The method includes obtaining a neural network topology and weights of a trained neural network. The method also includes transforming the neural network topology to an equivalent analog network of analog components including operational amplifiers and resistors. Each operational amplifier represents an analog neuron, and each resistor represents a connection between analog neurons. The method also includes computing a weight matrix for the equivalent analog network based on the weights of the trained neural network. The method also includes generating a resistance matrix for the weight matrix. The method also includes generating lithographic masks for fabricating a circuit implementing the equivalent analog network based on the resistance matrix. The method also includes fabricating the circuit based on the one or more lithographic masks using a lithographic process.

Abstract: Systems and methods are provided for analog hardware realization of neural networks. The method includes obtaining a neural network topology and weights of a trained neural network. The method also includes transforming the neural network topology to an equivalent analog network of analog components including operational amplifiers and resistors. Each operational amplifier represents an analog neuron of the equivalent analog network, and each resistor represents a connection between two analog neurons. The method also includes computing a weight matrix based on the weights of the trained neural network. The method also includes generating a resistance matrix for the weight matrix. The method also includes pruning the equivalent analog network to reduce the number of operational amplifiers or the resistors, based on the resistance matrix, to obtain an optimized analog network of analog components.

Abstract: Systems and techniques that facilitate data diversity visualization and/or quantification for machine learning models are provided. In various embodiments, a processor can access a first dataset and a second dataset, where a machine learning (ML) model is trained on the first dataset. In various instances, the processor can obtain a first set of latent activations generated by the ML model based on the first dataset, and a second set of latent activations generated by the ML model based on the second dataset. In various aspects, the processor can generate a first set of compressed data points based on the first set of latent activations, and a second set of compressed data points based on the second set of latent activations, via dimensionality reduction. In various instances, a diversity component can compute a diversity score based on the first set of compressed data points and second set of compressed data points.

Abstract: A virtual whiteboarding application is configured to interface with an issue tracking system to visualize and organize issues managed by the issue tracking system in a graphical user interface of the virtual whiteboarding application. Graphical elements corresponding to issues managed by the issue tracking system can be interacted with in the graphical user interface of the virtual whiteboarding application to relate issues in the issue tracking system and visualize relationships between issues managed by the issue tracking system.

Abstract: A method of improving usability and transparency of machine-learning aspects of applications providing various types of services is disclosed. Based on a request submitted through an administrative user interface, a data readiness check is performed on underlying data associated with the application. Based on a successful completion of the data readiness check, a configuration file is retrieved from an application server. The configuration file specifies a plurality of keys for generating a machine-learned model for the application. The machine-learned model is trained based on the plurality of keys specified in the configuration file. The machine-learned model is selected from a plurality of machine-learned models based on dry runs of the each of the plurality of models. The machine-learned model is activated with respect to the application. Scores are identified from the underlying data items based on the selected machine-learned model.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for generating an association between a graphic element and text for a messaging thread and providing an overlay including the graphic element based on providing a message to the messaging thread including the text. In particular, in one or more embodiments, the disclosed systems receive user selection of the text and the graphic element and provide the association between the text and the element to various client devices associated with the messaging thread. Accordingly, the disclosed systems provide instructions for each client device associated with the messaging thread to generate and provide an overlay for the messaging thread.

Abstract: A system and method for producing a video-based group conversation for viewing over a network. The system is configured to generate a display at a first user computer which includes a plurality of sections in a single screen display, including at least a preview section configured for displaying an indicator for each of one or more viewers who are viewing the video-based conversation, and a media player section configured for displaying a video signal including at least some of the video feeds. The creator of the video-based group conversation controls which users can participate in the video-based group conversation by enabling video feeds provided by those users to be displayed on the screen and viewable to other users who access the group conversation.

Abstract: A computer system includes memory configured to store a document database and a machine learning model. The document database includes multiple historical documents each having at least one version labeled as compliant and at least one version labeled as non-compliant. The system includes a creator user interface, a compliance user interface, an automated distribution module, and a model building module configured to train the machine learning model to classify a document according to a compliance score indicating a likelihood of document compliance with one or more compliance criteria. The system also includes an orchestrator module configured to receive the compliance score for the submitted document from the machine learning model, determine whether the compliance score is greater than or equal to a compliance score threshold, and supply the submitted document to the compliance user interface for transmission to the compliance team device when the compliance score is above a threshold.

Abstract: Cognitive aids for various aspects of an audio book may be provided to aid an end-user in following the book's characters and plot and to remind the end-user about prior aspects of the audio book. In some specific examples, the aid may be requested audibly, and the plot data and/or other data may be presented audibly.

Abstract: A differential information extraction unit 81 receives input of a function and a provisional solution of the function, and calculates a first-order differential and a second-order differential in the provisional solution of the input function. An end condition determination unit 82 determines whether to continue updating of a solution in the function, based on the provisional solution. A number-of-updates decision unit 83 decides a number of updates of the provisional solution, based on the provisional solution, the first-order differential, and the second-order differential. A provisional solution updating unit 84 repeatedly performs a process of updating the provisional solution as many times as the decided number of updates based on the first-order differential and the second-order differential, and outputs the updated provisional solution as an optimum solution.

Abstract: An on-image navigation guide can be generated utilizing coordinates of table data of a table in a document image. The on-image navigation guide enables a user to populate a data entry form with the table data by interacting directly with the document image. The data entry form is configured with the same number of columns as the table and the document image and the data entry form share the same data structure. The on-image navigation guide is presented on top of the document image. As a user navigates the document image, a focus point of the on-image navigation guide moves from table cell to table cell, providing an immediate visual feedback on what is captured in a corresponding data entry field. The on-image navigation guide includes an on-image context menu and a pin through which the user can make adjustments and/or corrections without leaving the document image.

Abstract: Method of generating a contextual action bar starts with processor causing an application icon associated with an application to be displayed by a display screen of a client device. Processor receives selection of application icon from a user and determines a context of client device. Context can comprise identification of application, type associated with application, or type of interface including application icon. Processor generates action bar based on the context of the client device, causes a first portion of display screen to display an application interface associated with the application, and causes a second portion to display the action bar that is associated with a messaging system. Other embodiments are also disclosed herein.

Abstract: Systems, methods, and products for auto tagging structured PDF documents that do not have accessibility tags. In one embodiment, structured PDF documents having accessibility tags are first parsed and analyzed to organize the visual components of the documents. The relationships of the identified objects to DOM elements (e.g., tags) are determined, and the objects and related DOM elements are stored in training files. The training files are used to train various classifiers. Untagged PDF documents are then parsed to identify included visual objects, and the classifiers are used to determine DOM elements that should be associated with visual objects identified in the untagged PDF documents. This information is used to construct a DOM structure corresponding to each untagged document. A new PDF is then generated corresponding to each untagged document using the generated DOM structure and visual object information.

Abstract: According to an embodiment, an electronic device may include a display, a memory, and at least one processor operatively coupled to the display and the memory. The memory stores instructions that are configured to, when executed, enable the at least one processor to display first stroke data of a first stroke trajectory, based on an input of a first touch-move after a first touch-down on the display, perform shape recognition on the first stroke trajectory, based on a size of the first stroke trajectory and a distance between a stroke start point of the first stroke trajectory and another point on the first stroke trajectory, convert the first stroke data into second stroke data corresponding to a recognized shape form, based on the shape recognition, and display the second stroke data on the display. Various other embodiments may be possible.

Abstract: It is possible to analyze what type of input/output mapping is mainly performed in the entire hidden layer of a neural network. A relationship analysis unit 30 calculates strength of a relationship of each combination of a dimension of the input data and a unit of the neural network and calculates strength of a relationship of each combination of the unit and a dimension of the output data. A role analysis unit 32 calculates a relationship between a prescribed number of types of roles and the unit and a relationship between the prescribed number of types of roles, the dimension of the input data, and the dimension of the output data on the basis of the strength of the relationship of each combination of the dimension of the input data and the unit of the neural network and the strength of the relationship of each combination of the unit and the dimension of the output data.

Abstract: Deep learning in the form of a deepfakes fabricator can be utilized to analyzed proposed documents, images, webpages etc. in order to ensure compliance with UI/UX standards. A rover component can scan and compartmentalize detected objects. A deepfakes fabricator can identify source datasets corresponding to detected objects and transform the objects so that they comply with standards stored in a web integrant. Resulting objects may be combined into an integrated image and an operational synthetic webpage may be created. Similarly, the disclosed deep-learning and other processes may operate in reverse and utilize an approved document, image, website, etc. in order automatically create (in whole or in part) a set of UI/UX standards.

Abstract: Systems, computer program products, and methods are described herein for automatically generating resource distributions. The present invention may be configured to receive a text-based instruction and parse, using a machine learning model, the text-based instruction to generate a structured resource distribution including predicted distribution elements. The present invention may be configured to generate, based on the structured resource distribution, a resource distribution. In some embodiments, the text-based instruction may include an email message, an SMS message, recorded speech converted to text, text input to a chat function, text recognized in an image, and/or the like.