Abstract: Systems and methods are provided for sorting cells with distinct cell designs for characterizing a library of the cell designs. The present disclosure uses randomized sorting rules associated with bins and pseudo-random numbers to counts the cells with measured fluorescence values in one of the bins. A mean fluorescence value for a cell design group may be determined based on a ratio of cell counts of the cells associated with the cell design group across the bins. Unlike the traditional histogram-based sorting that use a mean fluorescence value of a bin, the disclosed technology determines mean fluorescence values of cell design groups for characterizing libraries of the cell design group. Use of the mean fluorescence values with unbiased “sort-seq” and a de-multiplexed sequencing using the mean fluorescence values enables characterizing libraries of cell designs with improved accuracy over traditional use of discrete histograms.

Abstract: The discussion relates to virtual keyboard engagement. One example can define key volumes relating to keys of a virtual keyboard and detect finger movement of a user through individual key volumes. The example can detect parameter changes associated with detected finger movement through individual key volumes and build potential key sequences from detected parameter changes.

Abstract: A dual-screen computing device includes two separate displays that are coupled to an interconnecting hinge. A hinge detector detects movement or position of the hinge, and the positions of the displays may be determined based on the hinge movement or position. The positions of the displays relative to each other may then be used to determine which mode of operation the dual-screen computing device is operating (e.g., tent mode, open, closed, etc.). Additionally, the dual-screen computing device may include various sensors that detect different environmental, orientation, location, and device-specific information. Applications are configured to operate differently based on the mode of operation and, optionally, the sensor data detected by the sensors.

Abstract: The description relates to computing devices that employ steerable optics. One example includes a steering mechanism and a base substrate positioned relative to the steering mechanism. The example also includes an optical substrate positioned over the base substrate and an adhesive complex securing the optical substrate relative to the base substrate with multiple different types of adhesives.

Abstract: Techniques are described herein that are capable of providing source control of custom code in a backend-as-a-service (BaaS) module. For instance, the custom code may be automatically loaded dynamically (e.g., on-demand) in a runtime that is hosted by the BaaS module while the runtime is running to provide custom backend features that are defined by the custom code to an application in response to receipt of a request from the application. Source control of the custom code may be provided without providing source control of host code that defines the runtime.

Abstract: A method for single-handed microgesture input comprises receiving hand tracking data for a hand of a user. A set of microgesture targets that include software functions are assigned to positions along a length of a first finger. The received hand tracking data is analyzed by a gesture recognition machine. A location of a thumbtip of the hand of the user is determined relative to the positions along the first finger. Responsive to determining that the thumbtip is within a threshold distance of the first finger at a first position along the length of the first finger, a corresponding first microgesture target is designated for selection. Selection of the first microgesture target is enabled based on a duration the thumbtip is at the first position. Responsive to detecting a confirmation action, the corresponding microgesture target executes.

Abstract: Methods and systems related to memory reduction in a system by oversubscribing physical memory shared among compute entities are provided. A portion of the memory includes a combination of a portion of a first physical memory of a first type and a logical pooled memory associated with the system. A logical pooled memory controller is configured to: (1) track both a status of whether a page of the logical pooled memory allocated to any of the plurality of compute entities is a known-pattern page and a relationship between logical memory addresses and physical memory addresses associated with any allocated logical pooled memory, and (2) allow the write operation to write data to any available space in the second physical memory of the first type only up to an extent of physical memory that corresponds to the portion of the logical pooled memory previously allocated to the compute entity.

Abstract: Provided is a method, performed by a web browser, for identifying and displaying content, the method including receiving a webpage including text, comparing the text to titles of media content in a media content database, based on the comparison, identifying a subset of text as a match to a title of a matching media content, based on identifying the subset of text, modifying a display of the webpage by inserting a visual indicator into the webpage in proximity to the subset of text, or modifying the display of the subset of text, detecting an interaction of a user-controlled input with the inserted visual indicator or the modified display of the subset of text, and displaying a popup interface over a portion of the webpage, the popup interface displaying additional information about the matching media content and a selectable user-interface element to play the matching media content.

Abstract: An analysis application utilizes interactivity signals to generate relationships and promote content. One or more interactivity applications, such as a social networking application, are queried to retrieve interactivity signals. Interactivity signals include an interaction pattern that indicates a relationship between a user and relations of the user. A relationship graph is constructed based on the interactivity signals. Content associated with a user is promoted based on the relationship graph. A weight of the interactivity signals is adjusted to improve a ranking of the relationship graph and a ranking of the content.

Abstract: Aspects of this present disclosure relate to hybrid conference user interface. The hybrid conference interface provides an establishing shot before the meeting begins that places meeting attendees in a specific spatial arrangement, such as in specific seats around a conference table. Upon starting the conference, the hybrid user interface renders an appropriate perspective view of the meeting that is tailored to each attendee's perspective while also being spatially consistent for the entire group of attendees. Allowing attendees to place themselves where they want gives attendees a sense of physical space that helps them stay spatially oriented relative to the other people and resources in the room.

Abstract: Techniques for using online engagement footprints for video engagement prediction are provided. In one technique, events are received from multiple client devices, each event indicating a type of engagement of a video item from among multiple types of engagement. One or more machine learning techniques are used to train a prediction model that is based on the events and multiple features that includes the multiple types of engagement. In response to receiving a content request, multiple entity feature values are identified for a particular entity that is associated with the content request. Two or more of the entity feature values correspond to two or more of the types of engagement. A prediction is generated based on the entity feature values and the prediction model. The prediction is used to determine whether to select, from candidate content items, a particular content item that includes particular video.

Abstract: The syntax elements of a source code program used to represent the context of a focal method are selected based on a priority order. The selected syntax elements are input into a fixed-size context window that is used to train a neural transformer with attention model to learn to generate source code and used by the neural transformer model to generate source code. The context window contains prioritized sequences of tokens that extend beyond the target focus in order to provide a longer visibility back into the source code program for the model to learn predictive patterns. This gives the model a file-level context of the source code program without increasing the size of the context window.

Abstract: A near-eye display device comprises a pupil-expansion optic, first and second lasers, a drive circuit coupled operatively to the first and second lasers, a beam combiner, a spatial light modulator (SLM), and a computer. The first and second lasers are configured to emit in respective first and second wavelength bands. The beam combiner is configured to geometrically combine emission from the first and second lasers into a collimated beam. The SLM is configured to receive the collimated beam and to direct the emission in spatially modulated form to the pupil-expansion optic. The computer is configured to parse a digital image, trigger the emission from the first and second lasers by causing the drive circuit to drive current through the first and second lasers, and control the SLM such that the spatially modulated form of the emission projects an optical image corresponding to the digital image.

Abstract: Examples are disclosed that relate to image reprojection. One example provides a method, comprising receiving a first rendered image comprising content associated with a viewer reference frame, receiving a second rendered image comprising content associated with a manipulator reference frame, and reprojecting the first rendered image based on a head pose of a user to thereby produce a first reprojected image. The method further comprises reprojecting the second rendered image based on the head pose of the user and a pose of the manipulator to thereby produce a second reprojected image, and outputting the first reprojected image and the second reprojected image for display as a composited image.

Abstract: An electronic device detects a user presence in a first distance range from a user detection sensor of an electronic device. The electronic device communicates with multiple input components, setting the electronic device to a first device state based on detecting a user presence in the first distance range. The first device state applies a scanning rate to a first input component. The electronic device also detects a user presence in a second distance range from the user detection sensor of the electronic device while the electronic device is in the first device state. The electronic device sets the electronic device to a second device state, based on detecting a user presence in the second distance range while the electronic device is in the first device state. The second device state applies a different scanning rate to the first input component than in the first device state.

Abstract: For generating visual enhancement suggestions for source content, a system performs storing, in a data storage, a plurality of context data sets, each context data set including a set of visual enhancements and a context for selecting the set of visual enhancements; receiving the source content including source content data and source attribute data; providing, to an artificial intelligence (AI) engine, the received source content, the AI engine configured to select, based on the source content and the context data sets, a first set of visual enhancements and apply the selected first set of visual enhancements to the source content to generate a first visual enhancement suggestion for the source content; extracting, from the AI engine, the first visual enhancement suggestion; and causing the first visual enhancement suggestion to be displayed via a display of a user device.

Abstract: Techniques performed by a data processing system for training a Recurrent Neural Network Transducer (RNN-T) herein include encoder pretraining by training a neural network-based token classification model using first token-aligned training data representing a plurality of utterances, where each utterance is associated with a plurality of frames of audio data and tokens representing each utterance are aligned with frame boundaries of the plurality of audio frames; obtaining first cross-entropy (CE) criterion from the token classification model, wherein the CE criterion represent a divergence between expected outputs and reference outputs of the model; pretraining an encoder of an RNN-T based on the first CE criterion; and training the RNN-T with second training data after pretraining the encoder of the RNN-T. These techniques also include whole-network pre-training of the RNN-T.

Abstract: An automated natural dialogue system provides a combination of structure and flexibility to allow for ease of annotation of dialogues as well as learning and expanding the capabilities of the dialogue system based on natural language interactions.

Abstract: An automated tool analyzes source code repositories and web endpoints for unique characteristics that they both share in order to predict the likelihood that a particular source code repository contains source code files used in a web endpoint and to predict the likelihood that a web endpoint uses source code files of a particular repository. The unique characteristics are referred to as fingerprints and include unique combination of public-facing entities, unique tokens, and unique DOM characteristics.

Abstract: A system for extracting a key phrase from a document includes a neural key phrase extraction model (“BLING-KPE”) having a first layer to extract a word sequence from the document, a second layer to represent each word in the word sequence by ELMo embedding, position embedding, and visual features, and a third layer to concatenate the ELMo embedding, the position embedding, and the visual features to produce hybrid word embeddings. A convolutional transformer models the hybrid word embeddings to n-gram embeddings, and a feedforward layer converts the n-gram embeddings into a probability distribution over a set of n-grams and calculates a key phrase score of each n-gram. The neural key phrase extraction model is trained on annotated data based on a labeled loss function to compute cross entropy loss of the key phrase score of each n-gram as compared with a label from the annotated dataset.