Abstract: Techniques are described for managing inking events from remote inking devices during online meetings. For example, a remote inking device comprising a physical orientation sensor can be used during an online meeting to perform inking actions while shared screen content is being presented during the online meeting. To control the inking action, a stream of coordinate update events can be received from the remote inking device. The stream of coordinate update events represents physical movement of the remote inking device as indicated by the physical orientation sensor. The coordinate update events can be relayed to participant devices of the online meeting. The participant devices can then perform the inking action on local displays.

Abstract: Examples are disclosed that relate to the use of an optical data receiver comprising a multi-tap image sensor pixel for use in optical communications. The multi-tap pixel includes a photodetector and a plurality of taps. The optical data receiver further includes a controller comprising instructions executable for controlling the multi-tap pixel to, in a first period of time, perform a first integration on the photodetector and readout charge stored on a floating diffusion capacitor of a first tap in the plurality of taps using readout circuitry of the first tap. The controller further includes instructions executable for controlling the multi-tap pixel to, in a second period of time, perform a second integration on the photodetector and readout charge stored on a floating diffusion capacitor of a second tap in the plurality of taps using readout circuitry of the second tap.

Abstract: This document relates to searching for images using a query image. For example, an input image can be received and analyzed to identify objects, and then identification of two or more selected objects can be received. A query image can be derived from the input image, where the query image includes the selected objects. The query image can be used to perform a search to identify related images, which can be output in response to the query image.

Abstract: Examples of the present disclosure describe systems and methods for establishing a persistent connection between file attachments uploaded to cloud storage and associated electronic communications. An electronic communication including an attached file may be stored and presented in a communication application user interface that includes interactive functions, such as edit and save functions. Upon selection of an interactive function, a copy of the file may be uploaded to a cloud storage location and, in some examples, opened in a productivity application. A link to the file at the cloud storage location may be stored as a property of the file that is stored with the electronic communication. Resultantly, a connection may be maintained between the file at the cloud storage location and the electronic communication. Based on this connection, the copy of the file at the cloud storage location may be opened upon a subsequent selection of an interactive function.

Abstract: A method of training a neural network for detecting target features in images is described. The neural network is trained using a first data set that includes labeled images, where at least some of the labeled images having subjects with labeled features, including: dividing each of the labeled images of the first data set into a respective plurality of tiles, and generating, for each of the plurality of tiles, a plurality of feature anchors that indicate target features within the corresponding tile. Target features that correspond to the plurality of feature anchors are detected in a second data set of unlabeled images. Images of the second data set having target features that were not detected are labeled. A third data set that includes the first data set and the labeled images of the second data set is generated. The neural network is trained using the third data set.

Abstract: A computer-implemented method of generating a mixed reality workflow is described. The method comprises identifying a series of tasks and generating an input task-to-object-mapping by analyzing data that defines a process performed by a first user interacting with objects in a first location. The input task-to-object-mapping that maps each task from the series of tasks to an object used in the respective task. A task-specific non-spatial characteristic of each object in the input task-to-object-mapping is determined and used to map each object in the input task-to-object-mapping to a candidate object identified at a second location to generate an output task-to-object-mapping. The series of tasks, location data defining a position of each candidate object in the second location and output task-to-object-mapping are used to generate a mapped workflow which is then output to a device in the second location.

Abstract: A technique executes tasks using a data store of machine-trained models. The data store specifically includes a subset of encoder-type machine-trained models for converting input data items having different input data types into respective embeddings in a vector space, and a subset of decoder-type machine-trained models for converting embeddings in the same vector space into data items having respective different output data types. When executing a particular task that involves one or more data types, the technique selects one or more machine-trained models that match those data types. In some implementations, the technique provides a clipboard store for storing embeddings produced by the encoder-type machine-trained models and consumable by the decoder-type machine-trained models. The technique includes provisions for ensuring that any decoder-type machine-model is capable of processing embeddings produced by different versions of the encoder-type machine-trained models.

Abstract: A method for electronic accessory pairing includes capturing an image of an external environment via a camera communicatively coupled to a host computing device. The image of the external environment is analyzed to detect presence of an imaged electronic accessory. After determining that the host computing device is not presently paired with the imaged electronic accessory, a pairing is established between the host computing device and the imaged electronic accessory.

Abstract: Aspects of the present disclosure relate to systems and methods for performing targeted searching based on a user profile. In examples, a user profile including a user embedding may be retrieved based on the receipt of a user indication. The user embedding may be created based on one or more user interest. A plurality of document embeddings may be identified based on the user embedding, where each document embedding of the plurality of document embeddings is determined to be within a first distance of the user embedding. In examples, a ranking for each document embedding of the plurality of document embeddings may be generated, where the ranking for each document embedding of the plurality of document embeddings is based on the user embedding. At least one document may be recommend based on a ranking associated with a document embedding.

Abstract: Systems and methods are provided for generating visualization data associated with raw data using a machine learning model. For example, the machine learning model may automatically generate a set of candidate analytics and/or a scenario for visualizing the raw data based on summary data. Given the summary data and answers to prompts for visualizing data, the generated candidate analytics may reflect a context of the raw data as intended by the user. A visualization code scaffold according to a visualization specification may be used to generate programmatic output that corresponds to the candidate analytics, which may thus be used to generate a visualization accordingly. In some examples, an infographic may further be generated based on the visualization and a prompt using a diffusion model.

Abstract: A system for database reversion is described. The system comprises: a database engine configured to host an active database; a log engine configured to generate transaction logs for transactions affecting the active database; a backup engine configured to create a backup data structure to allow for database reversion; and a memory buffer separate from the active database. A page in the active database has an associated page timestamp indicating a most recent update of the page in the active database. The database engine is configured to flush an updated copy of a page in the memory buffer to the active database. The backup engine is configured to, prior to the flush, store an image of the page in the active database to the backup data structure when the page in the active database is older than the time value related to the creation time of the backup data structure.

Abstract: A data processing system includes a transaction bus, a console application in communication with the transaction bus, and a view predictor subsystem in communication with the transaction bus. The transaction bus receives, from a user application executing on a client device, a call for visual information to be provided to the user application. The view predictor subsystem determines a likelihood that the visual information will be viewable within a viewport of the user application, and a plurality of respective values for a plurality of sources of the visual information are computed based on the likelihood and a respective priority for each source. The console application provides to the transaction bus the set of potential sources of the visual information, and the transaction bus selects, based on the computed values, one of the potential sources of the visual information to be the result, which is provided to the user application.

Abstract: A technique iteratively updates model weights of a teacher model and a student model. In operation, the teacher model produces noisy original pseudo-labeled training examples from unlabeled training examples. The technique weights the original pseudo-labeled training examples based on validation information. The technique then updates model weights of the student model based on the weighted pseudo-labeled training examples. The validation information, which is used to weight the original pseudo-labeled training examples, is produced by selecting labeled training examples based on an uncertainty-based factor and a similarity-based factor. The uncertainty-based factor describes an extent to which the student model produces uncertain classification results for the set of labeled training examples. The similarity-based factor describes the similarity between the set of labeled training examples and the unlabeled training examples.

Abstract: A data processing system implements receiving a textual context inserted into a user interface element; receiving an indicator inserted into the user interface element after the textual context, the indicator indicating a desire to tag a topic from a plurality of topics included in a knowledge base; receiving one or more textual character inserted into the user interface element after the indicator; encoding, using a machine-learning (ML) model, the received textual context to generate at least one representation reflecting one or more meanings of the received textual context; decoding, using the ML model, the at least one representation to generate a plurality of tokens in response to the one or more meanings of the received textual context, the plurality of tokens corresponding with the at least one textual character and at least one of the topics of the plurality of topics; identifying one or more topics from the plurality of topics as recommended topics; and providing the identified recommended topics for

Abstract: Rectangular equidistant projection is applied to coordinates for path fragments of a plurality of fiber optic routes. A bounding box is generated for each path fragment. For each cell of a grid of cells where its bounding box overlaps with a bounding box created by the cell, each path fragment is placed into the cell of the global grid. For any cell with more than one path fragment, each path fragment inside the cell is compared to every other path fragment in that cell. For any path fragment with overlapping bounding boxes, the length of the overlap is saved. A table of cumulative overlap distances between path fragments is generated to identify overlapping segments and cumulative lengths of the overlap distances.

Abstract: A computing device is configured to perform acts that include displaying a graphical user interface (GUI) of an operating system installed on the computing device, where a taskbar of the GUI includes a search box. The search box includes: 1) a text entry field that is configured to receive textual input; and 2) a graphical icon. The acts also include detecting a selection of the text entry field and presenting a first graphical pane on the display in response to detecting the selection of the text entry field, where the first graphical pane includes first content. The acts also include detecting a selection of the graphical icon. The acts further include presenting a second graphical pane on the display in response to detecting the selection of the graphical icon, where the second graphical pane comprises second content that is different from the first content.

Abstract: A multi-model data store comprises an enrichment catalog and an enrichment memory graph. Raw data from a source raw datastore is enriched based on an enrichment function. Enriched data is stored in graph nodes with edge associations indicating the enrichment type and confidence level. The enriched data is also stored in the enrichment catalog with full breadcrumb provenance and lineage identifying, with cataloged enrichment steps, down to a record and field in the source raw datastore. Additional enriched values and associations between graph nodes can be entered in the graph by a user. Queries to the multi-model database begin in the graph and continue to the enrichment catalog. Query results include the full sequential chain of enrichments to source raw data records and fields, and may include dynamically generated virtual enrichment catalog records that link search values not found in the enrichment catalog to enrichment catalog records via the graph.

Abstract: The present disclosure relates to processing operations that execute image classification training for domain-specific traffic, where training operations are entirely compliant with data privacy regulations and policies. Image classification model training, as described herein, is configured to classify meaningful image categories in domain-specific scenarios where there is unknown data traffic and strict data compliance requirements that result in privacy-limited image data sets. Iterative image classification training satisfies data compliance requirements through a combination of online image classification training and offline image classification training. This results in tuned image recognition classifiers that have improved accuracy and efficiency over general image recognition classifiers when working with domain-specific data traffic. One or more image recognition classifiers are independently trained and tuned to detect an image class for image classification.

Abstract: A thermal management system includes a server computer and a shell defining an immersion chamber. The shell contacts the server computer along a bottom side and lateral walls of the chamber, and the immersion chamber has an opening on a top side. The shell is connected to the server computer and a portion of the server computer defines at least a portion of the immersion chamber.

Abstract: Systems, methods, and software are disclosed herein that support enhanced code linting. In an implementation, an application hosts a content object that is the subject of one or more scripts for automating tasks. A linter obtains an initial state of the content object from the application and generates a model that reflects the object's state. The linter also performs an analysis of the code statements with respect to the model. The linter may then enable or otherwise cause one or more results of the analysis to be displayed in a user interface to a code editor used for interacting with the scripts.