Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for generating a visual media collection for a social networking account and provide access to (or distribute) images, videos, or other visual media items from the visual media collection separate from social networking posts uncategorized within such a collection for the social networking account. For example, based on follow requests specific to a visual media collection, the disclosed systems can further distribute visual media items in collections posts from a particular visual media collection differing from other visual media collections and from social networking posts uncategorized within such a collection of a social networking account.

Abstract: In one embodiment, a method includes training a baseline machine-learning model based on a neural network comprising a plurality of stages, wherein each stage comprises a plurality of neural blocks, accessing a plurality of training samples comprising a plurality of content objects, respectively, determining one or more non-local operations, wherein each non-local operation is based on one or more pairwise functions and one or more unary functions, generating one or more non-local blocks based on the plurality of training samples and the one or more non-local operations, determining a stage from the plurality of stages of the neural network, and training a non-local machine-learning model by inserting each of the one or more non-local blocks in between at least two of the plurality of neural blocks in the determined stage of the neural network.

Abstract: The present disclosure is directed toward systems and methods for optimizing video conferences. For instance, systems and methods described herein optimize both the transmission and display of one or more video conference data streams. Systems and methods described herein optimize the transmission and display of one or more video conference data streams by identifying a context associated with the one or more video conference data streams and optimizing the one or more video conference data streams based on the identified context.

Abstract: Systems, methods, and non-transitory computer-readable media can determine a user embedding for a user of a content provider based on one or more machine learning models. A job embedding for a job posting published through the content provider can be determined based on one or more machine learning models. A determination can be made of a threshold level of similarity between the user embedding and the job embedding. The job posting can be provided as a recommendation to the user through the content provider.

Abstract: One or more embodiments of the disclosure provide systems and methods for providing media presentations to users of a media presentation system. A media presentation generally includes a one or more media streams provided by one or more capturing users of the media presentation system. In one or more embodiments, a user of the media presentation system may share a media presentation with a viewing user. The media presentation system can provide a number of features to produce a media stream within a media presentation.

Abstract: In one embodiment, a method includes receiving a media content for delivery to a first user, wherein the media content is associated with a priority level, accessing context information associated with the first user, wherein the context information is determined based on one or more multimodal signals from a client device associated with the first user, determining a delivery level for the media content based on the priority level of the media content and the context information, and instructing the client device to present a notification associated with the media content to the first user, wherein the notification is rendered in one or more modalities determined based on the delivery level for the media content.

Abstract: Various features of a high performance hardware scaler are disclosed herein. In some embodiments, a hardware scaler comprises a first processing unit configured to perform preparation and scaling operations and a second processing unit configured to perform preparation and scaling operations. The first processing unit and the second processing unit alternatively switch between performing preparation and scaling operations when processing a current input pixel block such that the first processing unit performs scaling operations while the second processing unit performs preparation operations and the second processing unit performs scaling operations while the first processing unit performs preparation operations.

Abstract: A system comprises a memory storage configured to store at least a portion of a frame of a video and a hardware motion estimation search processing unit configured to perform at least a portion of a motion estimation search for the video for a plurality of different block sizes. The hardware motion estimation search processing unit is configured to perform the motion estimation search using a plurality of source sub-blocks of a first block size to determine a first type of comparison evaluation values for the first block size. A combination of values included in the first type of comparison evaluation values is utilized to determine at least one second type of comparison evaluation values for a second block size, wherein the second block size is larger than the first block size.

Abstract: Two streams of data items are received. A first estimated processing time for the first stream of data items and a second estimated processing time for the second stream of data items are determined. Data items of the first stream and data items of the second stream are dynamically maintained in a first buffer and a second buffer respectively. The data items of the second stream maintained in the second buffer have associated event times that are within a first join window based on the first estimated processing time for the first stream. A selected data item of the first stream maintained in the first buffer is joined with one or more data items of the second stream maintained in the second buffer that have associated event times that are within a second join window based on an event time associated with the selected data item of the first stream.

Abstract: A request to perform a compute task is received. A plurality of compute processor resources eligible to perform the compute task is identified, wherein the plurality of compute processor resources includes two or more of the following: a field-programmable gate array, an application-specific integrated circuit, a graphics processing unit, or a central processing unit. Based on an optimization metric, one of the compute processor resources is dynamically selected to perform the compute task.

Abstract: Various client devices include displays and one or more image capture devices configured to capture video data. Different users of an online system may authorize client devices to exchange information captured by their respective image capture devices. Additionally, a client device modifies captured video data based on users identified in the video data. For example, the client device changes parameters of the image capture device to more prominently display a user identified in the video data and may further change parameters of the image capture device based on gestures or movement of the user identified in the video data. The client device may apply multiple models to captured video data to modify the captured video data or subsequent capturing of video data by the image capture device.

Abstract: Techniques are described for a social networking system to enable multiple users to collaboratively create, publish, receive attribution, and/or edit content. In some cases, the social networking system may be configured to distribute the collaborative content and feedback associated therewith based on metrics and relationships between collaborators of the collaborative content and users of the social networking system.

Abstract: A mobile client device is localized based on a captured image by identifying where the client device is located from a set of known locations. The set of known locations is associated with a set of regions, where each region is associated with a set of key frames representing the important features of the region. Latent vectors and keypoints are calculated for each of the key frames and an image captured by the client device. The system compares the latent vectors of the captured image to the latent vectors associated with the regions to determine a subset of similar regions. The system compares the keypoints of the captured image to the keypoints associated with the regions in the subset to determine a best match. This determined location is considered the region of the client device and may be used with other localization information to maintain localization of the client device.

Abstract: Data characterizing historical skills-based gaming metrics for a first user and historical skills-based gaming metrics for at least one second user is accessed. Using the accessed data and a set of rules, a targeted advertisement to present to the first user is determined. The targeted advertisement specifies at least one skills-based game and a characterization of the at least one second user's historical skills-based gaming metrics. The targeted advertisement is generated. Data characterizing the targeted advertisement is provided. Related apparatus, systems, techniques, and articles are also described.

Abstract: In one embodiment, a method includes generating audio segments from a speech signal, generating latent representations that respectively correspond to the audio segments, the latent representations comprising a first subset and a second subset, generating quantized representations that respectively correspond to the latent representations, masking the second subset of the latent representations, using a machine-learning model to process the first subset of the latent representations and the masked second subset of the latent representations to generate contextualized representations that respectively correspond to the latent representations, pre-training the machine-learning model based on comparisons between (1) a subset of the contextualized representations that respectively correspond to the masked second subset of the latent representations and (2) a subset of the quantized representations that respectively correspond to the masked second subset of the latent representations, and training the pre-trained

Abstract: Techniques for messaging bot controls based on machine-learning user intent detection are described. In one embodiment, an apparatus may comprise a message queue monitoring component operative to monitor a messaging interaction, the messaging interaction exchanged via a messaging system, the messaging interaction involving at least one client device; an interaction processing component operative to determine a user intent for the messaging interaction; and a bot management component operative to determine a messaging bot options configuration for the client device based on the user intent; and send the messaging bot options configuration to the client device. Other embodiments are described and claimed.

Abstract: The disclosed method may include detecting motion of a user, estimating, for the detected motion of the user, effort expended by the user in performing the motion, determining, based on the detected motion and the estimation of expended effort, a haptic profile for conveying to the user a physical sensation of expending the effort, and simulating a sensation of expending the effort by executing the haptic profile in at least one haptic device that is worn by the user. Various other methods, systems, and/or computer-readable media are also disclosed.

Abstract: Methods, systems, and non-transitory computer readable storage media are disclosed for two-phase encoding a digital video based on downsampling distortion of the digital video and a constant rate factor transition threshold. For example, the disclosed system can determine a downsampling distortion indicating a measure of distortion resulting from downsampling an input digital video. The disclosed systems can utilize the downsampling distortion to determine a constant rate factor transition threshold for selecting sets of encoding parameters. For example, the disclosed systems can select a first set of encoding parameters below the constant rate factor transition threshold and a second set of encoding parameters at or above the constant rate factor transition threshold. Additionally, the disclosed systems can generate first and second sets of digital video encodings of the input digital video by utilizing the first and second sets of encoding parameters, respectively.

Abstract: Systems and methods for screenless computerized social-media access may include (1) producing, via an audio speaker that is communicatively coupled to a computing device, a computer-generated verbal description of a social-media post provided via a social-media application, (2) detecting, via a microphone that is communicatively coupled to the computing device, an audible response to the social-media post from a user of the computing device, and (3) digitally responding to the social-media post in accordance with the detected audible response. Various other methods, systems, and computer-readable media are also disclosed.