Abstract: A processor system comprises two groups of registers and a hardware channel convolution processor unit. The first group of registers is configured to store data elements of channels of a portion of a convolution data matrix. Each register stores at least one data element from each channel. The second group of registers is configured to store data elements of convolution weight matrices including a separate matrix for each channel. Each register stores at least one data element from each matrix. The hardware channel convolution processor unit is configured to multiply each data element in a first and second portion of the first group of registers with a corresponding data element in the second group of registers to determine corresponding multiplication results and sum together the multiplication results for each specific channel to determine two corresponding channel convolution result data elements in a corresponding channel convolution result matrix.

Abstract: A first group of elements is element-wise multiplied with a second group of elements using a plurality of multipliers belonging to a matrix multiplication hardware unit. Results of the plurality of multipliers are added together using a hierarchical tree of adders belonging to the matrix multiplication hardware unit and a final result of the hierarchical tree of adders or any of a plurality of intermediate results of the hierarchical tree of adders is selectively provided for use in determining an output result matrix.

Abstract: Systems, methods, and non-transitory computer-readable media can collect a set of training videos as training data, wherein the set of training videos are labeled with one or more labels based on one or more video quality metrics associated with an evaluation objective. A machine learning model is trained based on the training data. A video to be evaluated is received. The video is assigned to a first video quality category of a plurality of video quality categories based on the machine learning model.

Abstract: According to examples, a system for generating and delivering enhanced content utilizing remote rendering and data streaming is described. The system may include a processor and a memory storing instructions. The processor, when executing the instructions, may cause the system to access a first data store with first information and a second data store with second information and align the first information with the second information to generate an aligned set. The processor, when executing the instructions, may then perform a computation on one or more identifiers utilizing the generated aligned set and reveal a differentially private output to one or more receiving parties.

Abstract: Techniques for efficient messaging client communication are described. In one embodiment, an apparatus may comprise a client front-end component of a messaging server operative to receive a protocol connect packet from a messaging client, the protocol connect packet initiating a client authentication process with the messaging client; and receive a message package from the messaging client, the message package received during the client authentication process; a client authentication component operative to perform the client authentication process with the messaging client; and a client message queueing component operative to queue the message package in an authorization-holding message queue in response to the message package being received during the client authentication process. Other embodiments are described and claimed.

Abstract: Systems, methods, and non-transitory computer-readable media can obtain respective instances of verification information can be obtained from computing systems associated with a plurality of third-party vendors, wherein a computing system associated with a third-party vendor determines verification information based at least in part on an analysis of an identification document provided by a user to be authenticated. The respective instances of verification information can be interpreted, wherein interpreting an instance of verification information comprises assigning respective states to attribute fields included in the verification information. A set of interpreted instances of verification information can be evaluated in combination to determine whether to authenticate the user, wherein the set includes an interpreted instance of verification information from each of the plurality of third-party vendors.

Abstract: A computing device receives a message including a request for a recommendation. A representation of a hypothetical ideal recommendation to provide in response to the message is determined based on the message content. Data regarding entities that are potential recommendations are retrieved from a data store, the data regarding each entity including a representation of the entity (e.g., a vector) derived from factual information about the entity and opinions of other users of the entity. Ranking scores are determined for at least a subset of the entities based on the difference between the entity representations and the representation of the hypothetical ideal recommendation. An entity to recommend is selected based on the ranking scores and a reply to the message is sent that identifies the selected entity.

Abstract: Systems and methods for co-present digital messaging may include (1) providing a group of users with access to a digital messaging thread that enables the users to exchange messages, (2) detecting a co-present moment in which two or more of the users are accessing the digital messaging thread at the same time, and (3) displaying a digital streak indicator proximate each message posted to the digital messaging thread during the co-present moment to visually identify the messages posted to the digital messaging thread during the co-present moment. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: A disclosed computer-implemented method may include determining that a user of a social networking system is a participant in a dating service of the social networking system. The method may also include identifying a set of interest-based communities of the social networking system of which the user is a member and that include an activated community-based dating feature. The method may also include selecting a set of additional participants in the dating service. Each additional participant included in the set of additional participants may be an additional member of at least one interest-based community included in the set of interest-based communities. The method may also include presenting, within a dating interface of the social networking system, dating information associated with at least a portion of the set of additional participants. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: The disclosed computer-implemented method may include compiling a neural network, and the compiling may include organizing an interconnected set of nodes in a series of layers, and for each node in each layer, assigning an associated activation of a plurality of activations. Each activation may output a respective tensor of a plurality of tensors. The compiling may also include allocating memory for the activations by determining a respective memory size for each activation, and based on the respective memory size for each activation, assigning a memory block in the neural network to the activation. The method may also include, after the allocating the memory for the activations, accessing the memory blocks to perform the plurality of activations and thereby execute the neural network. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: According to examples, a system for providing dynamic digital content may include a processor and a memory storing instructions. The processor, when executing the instructions, may cause the system to receive a plurality of data feeds. The processor may further analyze the data feeds to identify values for parameterized variables. A plurality of deep learning (DL) models can be trained to obtain product attribute data from the data feeds. The processor may then identify rules or triggers based on the values of the parameterized variables. The rules and/or triggers cause the processor to dynamically generate or select digital content and transmit the digital content to user communication devices of selected audience.

Abstract: In one embodiment, a method includes receiving SRS received from a plurality of UEs associated with the base station from an RU associated with the base station, estimating strengths or signal-to-noise ratios (SNRs) for pre-determined beams for each of the plurality of UEs based on the received SRS, selecting a subset of the plurality of UEs to which downlink data is to be transmitted for a RBG in a TTI based on the estimated strengths or SNRs of the pre-determined beams for the plurality of UEs, computing a precoding matrix for the RBG based on the selected subset, preparing multi-layered UE data for the RBG based on the selected subset and the computed precoding matrix, sending the multi-layered UE data and the precoding matrix for the RBG to the RU, where the RU transmits pre-coded multi-layered UE data to the UEs in the subset using MIMO technologies.

Abstract: In some examples, a device includes a fluid lens having a substrate and a membrane connected to the substrate using a flexure. The fluid lens may include a fluid located within a cavity at least partially defined by the substrate and the membrane. The flexure may include a membrane attachment and an elastic element. The device may be configured so that a displacement of the membrane attachment adjusts a profile of the membrane, and may induce a compression of at least a portion of the elastic member. In some examples, a flexure may include a plurality of elastic elements, which may be attached to the substrate (e.g., through a flexure support), and a rigid element, that may include or be connected to the membrane attachment. Example devices include head-mounted devices, such as augmented reality or virtual reality devices.

Abstract: Introduced are systems and methods that enable modification of logs in multiple off-line databases. Multiple off-line devices can mistakenly associate different respondents with the same identification (ID) unique to the system. When the multiple off-line devices synchronize with each other, or synchronize with a server hosting the central database, the software running on the off-line devices, or on the server detects that the modified logs come from different respondents, and the software assigns two different IDs unique to system to the logs. In another embodiment, multiple off-line devices can mistakenly associate the same respondent with two different IDs unique to the system. When the multiple off-line devices synchronize with each other or with the server, the software running on the off-line devices, or the server detects that the modified logs come from the same respondent, and the software assigns the logs to the same ID unique to system.

Abstract: A social networking system provides content items to a user via a feed that may include one or more sponsored content items. Multiple sponsored content items may be included in a set that is presented in the feed via a scrollable content unit that presents a sponsored content item from the set and presents additional sponsored content items from the set when user interaction is received. To place sponsored content items in the feed, the social networking system scores a set of sponsored content items based on prior user interactions with content presented via scrollable content units and a bid amount of a sponsored content item in the set. The set of sponsored content items is ranked among other sponsored content items based on its score. If the set of sponsored content items is selected for inclusion in the feed, the social networking system orders the sponsored content items in the set for presentation via the scrollable content unit.

Abstract: Systems, methods, and non-transitory computer-readable media can be configured to acquire at least one instance of positive training data based at least in part on at least one source. A set of supplemental positive training data can be generated based at least in part on the at least one instance of positive training data. A machine learning model can be trained to identify inappropriate material based at least in part on the set of supplemental positive training data.