Abstract: The present disclosure relates to systems, methods, and computer readable media that utilize a trained multi-modal combination model for content and text-based evaluation and distribution of digital video content to client devices. For example, systems described herein include training and/or utilizing a combination of trained visual and text-based prediction models to determine predicted performance metrics for a digital video. The systems described herein can further utilize a multi-modal combination model to determine a combined performance metric that considers both visual and textual performance metrics of the digital video. The systems described herein can further select one or more digital videos for distribution to one or more client devices based on combined performance metrics associated with the digital videos.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating digital video summaries based on analyzing a digital video utilizing a relevancy neural network, an aesthetic neural network, and/or a generative neural network. For example, the disclosed systems can utilize an aesthetics neural network to determine aesthetics scores for frames of a digital video and a relevancy neural network to generate importance scores for frames of the digital video. Utilizing the aesthetic scores and relevancy scores, the disclosed systems can select a subset of frames and apply a generative reconstructor neural network to create a digital video reconstruction. By comparing the digital video reconstruction and the original digital video, the disclosed systems can accurately identify representative frames and flexibly generate a variety of different digital video summaries.

Abstract: This disclosure relates to methods, non-transitory computer readable media, and systems that determine multiple personas corresponding to a user account for digital content and train a persona classifier to predict a given persona (from among the multiple personas) for content requests associated with the user account. By using the persona classifier, the disclosed methods, non-transitory computer readable media, and systems accurately detect a given persona for a content request upon initiation of the request. Based on determining the given persona, in some implementations, the methods, non-transitory computer readable media, and systems generate a digital-content recommendation for presentation on a client device associated with the user account.

Abstract: In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus may be a BMC. The BMC stores an extended SEL record entry in a data repository. The extended SEL record entry including a locator to a data collection stored external to the data repository. The BMC further retrieves the data collection based on the locator in response to receiving a request-to-retrieve-an-SEL-record.

Abstract: Techniques are disclosed for training of factorization machines (FMs) using a streaming mode alternating least squares (ALS) optimization. A methodology implementing the techniques according to an embodiment includes receiving a datapoint that includes a feature vector and an associated target value. The feature vector includes user identification, subject matter identification, and a context. The target value identifies an opinion of the user relative to the subject matter. The method further includes applying an FM to the feature vector to generate an estimate of the target value, and updating parameters of the FM for training of the FM. The parameter update is based on application of a streaming mode ALS optimization to: the datapoint; the estimate of the target value; and to an updated summation of intermediate calculated terms generated by application of the streaming mode ALS optimization to previously received datapoints associated with prior parameter updates of the FM.

Abstract: Techniques are disclosed for automatically assigning weights to attributes of media content based in part on how many users actually viewed or listened to the content, as well as how many users “liked” or otherwise indicated a preference for the content. The content items can be any type of audio or visual media content, such as songs, videos, or movies, as well as written content, such as books, articles, journals, advertisements, or magazines. A first similarity score is determined based on a similarity between user preferences for content items. A second similarity score is determined based on a similarity between one or more common attributes of the content items. These attributes are assigned ratings that represent the number of users who consumed the corresponding content. Next, weights are assigned to each of the attributes based on the first and second similarity scores using, for example, linear equation regression techniques.

Abstract: Techniques are disclosed for the improvement of vector quantization (VQ) codebook generation. The improved codebooks may be used for compression in cloud-based video applications. VQ achieves compression by vectorizing input video streams, matching those vectors to codebook vector entries, and replacing them with indexes of the matched codebook vectors along with residual vectors to represent the difference between the input stream vector and the codebook vector. The combination of index and residual is generally smaller than the input stream vector which they collectively encode, thus providing compression. The improved codebook may be generated from training video streams by grouping together similar types of data (e.g., image data, motion data, control data) from the video stream to generate longer vectors having higher dimensions and greater structure. This improves the ability of VQ to remove redundancy and thus increase compression efficiency.

Abstract: Embodiments of the present invention relate to facilitating selection of video advertisements for presentation in association with a video. In embodiments, advertisement quality associated with various video advertisements is referenced. Generally, the advertisement quality indicates a probability a viewer will continue viewing a portion of a video following presentation of the video advertisement presented in association with the video. The advertisement quality associated with the video advertisements is used to select one or more video advertisements for presentation along with the video. An indication of the selected video advertisements can be provided for integration with the video to present to the viewer.

Abstract: A system includes a computing device and a server management controller configured to be connected to the computing device. The server management controller includes a processor, an interface in communication with multiple sensors disposed in the computing device, and a non-volatile memory storing computer executable code. The code, when executed at the processor, is configured to: assign each of the sensors with a logical unit number (LUN) and a sensor number; receive a first message from a first sensor of the sensors via the interface, where the first message has a first LUN field storing the assigned LUN and a first sensor number field storing the assigned sensor number of the first sensor; process the received first message; and determine the first sensor as a source of the first message based on the LUN and the sensor number stored in the first message.

Abstract: Embodiments of the present invention relate to facilitating determining video advertisement quality. In embodiments, video advertisement data associated with a video advertisement or presentation of the video advertisement in association with a video is referenced. Such video advertisement data can indicate, for example, an advertisement characteristic, a video characteristic, a video session characteristic, and/or a viewer characteristic. The video advertisement data can be used to determine an advertisement quality associated with the video advertisement. The advertisement quality generally indicates a probability a viewer will continue viewing a portion of the video following presentation of the video advertisement presented in association with the video.

Abstract: Embodiments of the present disclosure provide systems, methods, and computer storage media for mitigating latencies associated with the encoding of digital assets. Instead of waiting for codebook generation to complete in order to encode a digital asset for storage, embodiments described herein describe a shifting codebook generation and employment technique that significantly mitigates any latencies typically associated with encoding schemes. As a digital asset is received, a single codebook is trained based on each portion of the digital asset, or in some instances along with each portion of other digital assets being received. The single codebook is employed to encode subsequent portion(s) of the digital asset as it is received. The process continues until an end of the digital asset is reached or another command to terminate the encoding process is received. To encode an initial portion of the digital asset, a bootstrap codebook can be employed.

Abstract: Embodiments of the present disclosure provide systems, methods, and computer storage media for optimizing computing resources generally associated with cloud-based media services. Instead of decoding digital assets on-premises to stream to a remote client device, an encoded asset can be streamed to the remote client device. A codebook employable for decoding the encoded asset can be embedded into the stream transmitted to the remote client device, so that the remote client device can extract the embedded codebook, and employ the extracted codebook to decode the encoded asset locally. In this way, not only are processing resources associated with on-premises decoding eliminated, but on-premises storage of codebooks can be significantly reduced, while expensive bandwidth is freed up by virtue of transmitting a smaller quantity of data from the cloud to the remote client device.

Abstract: Embodiments of the present disclosure provide systems, methods, and computer storage media for mitigating delays typically experienced when training codebooks during the encoding process. Instead of training a codebook based on a single digital asset, multiple digital assets determined to have asset characteristics in common can be grouped together to form a group of digital assets, from which a single codebook can be trained. The group of digital assets together form a codebook training set, such that each digital asset therein can be analyzed, in parallel, to expeditiously train a single codebook. A codebook trained in this manner can be employed to encode other digital assets sharing the asset characteristics as those in the codebook training set.

Abstract: Techniques are disclosed for the improvement of vector quantization (VQ) codebook generation. The improved codebooks may be used for compression in cloud-based video applications. VQ achieves compression by vectorizing input video streams, matching those vectors to codebook vector entries, and replacing them with indexes of the matched codebook vectors along with residual vectors to represent the difference between the input stream vector and the codebook vector. The combination of index and residual is generally smaller than the input stream vector which they collectively encode, thus providing compression. The improved codebook may be generated from training video streams by grouping together similar types of data (e.g., image data, motion data, control data) from the video stream to generate longer vectors having higher dimensions and greater structure. This improves the ability of VQ to remove redundancy and thus increase compression efficiency.

Abstract: Various embodiments describe data compression that implements vector quantization. A computer system generates a codebook for the vector quantization by iteratively clustering vectors representative of data that should be compressed. The iterative clustering uses geometric reasoning to avoid distance computations between vectors as appropriate, thereby reducing the latency associated with generating the codebook. Further, the system encodes the vectors based on the codebook. To do so, the computer system generates hashes of the vectors by applying locality sensitive hashing to these vectors. The hashes are compared and matched with hashes of codebook vectors. The computer system represents the vectors based on the matched codebook vectors.

Abstract: Various embodiments describe a dynamic reconfiguration of a media processing system to optimize a latency performance. In an example, a computer system accesses a current latency performance of the media processing system. The latency is associated with performing a codec process on a current configuration of the media processing system. The current configuration includes virtual machines. The computer system estimates, based on the current latency performance and on historical latency performances associated with the current configuration, a next latency performance of the media processing system. The computer system also identifies, from potential configurations, an updated configuration of the media processing system based on a difference between the next latency performance and a target latency performance and on historical performances associated with the potential configurations. The updated configuration specifies an additional number of virtual machines associated with hosting the codec process.

Abstract: In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus may be a service processor. The service processor emulates a frame buffer. The service processor determines whether the service processor is configured to generate display signals to be sent to a display based on images from the service processor or based on images from a host of the service processor. When the service processor is configured to generate the display signals based on the images from the service processor, the service processor generates a first image based on serial data received from the host through a serial port, writes the first image to the emulated frame buffer, and generates a first display signal to be sent to the display based on the first image from the emulated frame buffer.

Abstract: Various embodiments describe data compression that implements vector quantization. A computer system generates a codebook for the vector quantization by iteratively clustering vectors representative of data that should be compressed. The iterative clustering uses geometric reasoning to avoid distance computations between vectors as appropriate, thereby reducing the latency associated with generating the codebook. Further, the system encodes the vectors based on the codebook. To do so, the computer system generates hashes of the vectors by applying locality sensitive hashing to these vectors. The hashes are compared and matched with hashes of codebook vectors. The computer system represents the vectors based on the matched codebook vectors.

Abstract: Various embodiments describe a dynamic reconfiguration of a media processing system to optimize a latency performance. In an example, a computer system accesses a current latency performance of the media processing system. The latency is associated with performing a codec process on a current configuration of the media processing system. The current configuration includes virtual machines. The computer system estimates, based on the current latency performance and on historical latency performances associated with the current configuration, a next latency performance of the media processing system. The computer system also identifies, from potential configurations, an updated configuration of the media processing system based on a difference between the next latency performance and a target latency performance and on historical performances associated with the potential configurations. The updated configuration specifies an additional number of virtual machines associated with hosting the codec process.

Abstract: A buffer model in an HTTP streaming client may include receiving a first content fragment of a first content stream in response to a first HTTP request. It may also include receiving a second content fragment of a second content stream in response to a second HTTP request. The buffer model may further include storing the first and second content fragments in first and second buffers of a plurality of configurable buffers. The first and second content fragments may be multiplexed into a third buffer of the plurality of buffers. The multiplexed first and second content fragments may be stored in a fourth buffer of the plurality of buffers for playback. The buffer model may be implemented by an application. The buffers may be designed based on one or more constraints.