Abstract: Apparatus and methods relate to receiving an input image comprising an array of pixels, wherein the input image is associated with a first characteristic; applying a neural network to transform the input image to an output image associated with a second characteristic by generating, by an encoder and for each pixel of the array of pixels of the input image, an encoded pixel, providing, to a decoder, the array of encoded pixels, applying, by the decoder, axial attention to decode a given pixel, wherein the axial attention comprises a row attention or a column attention applied to one or more previously decoded pixels in rows or columns preceding a row or column associated with the given pixel, wherein the row or column attention mixes information within a respective row or column, and maintains independence between respective different rows or different columns; and generating, by the neural network, the output image.

Abstract: Systems, devices, circuitries, and methods are disclosed for identifying, within a call instruction, context registers for storing prior to a jump to another subroutine. In one example, a method includes receiving, while executing a first subroutine, a call instruction that includes a first opcode and identifies a first target address, wherein the first target address stores instructions for performing a second subroutine. A first set of context registers identified by the call instruction is determined and the content of the first set of context registers is stored in first memory allocated for context storage for the first subroutine prior to executing the instruction stored in the first target address.

Abstract: The techniques disclosed herein enable systems to train machine learning models using benign augmentation to enabled resistance various data poisoning attacks. This is achieved by first training a machine learning model using an initial dataset that is trustworthy and originates from a known source. The initial dataset is then modified to include known attack triggers such as syntactic paraphrasing to generate an augmented dataset. The augmented dataset is then used to train a robust machine learning model based using the initially trained machine learning model. The resultant robust machine learning model is then enabled to detect and resist attacks captured by the augmented dataset. The robust machine learning model can be retrained using an untrusted dataset that includes various compromised inputs in conjunction with the augmented dataset. Retraining results in an updated robust machine learning model that can learn and resist various data poisoning attacks on the fly.

Abstract: Works in the literature fail to leverage embedding access patterns and memory units' access/storage capabilities, which when combined can yield high-speed heterogeneous systems by dynamically re-organizing embedding tables partitions across hardware during inference. A method and system for optimal deployment of embeddings tables across heterogeneous memory architecture for high-speed recommendations inference is disclosed, which dynamically partitions and organizes embedding tables across fast memory architectures to reduce access time. Partitions are chosen to take advantage of the past access patterns of those tables to ensure that frequently accessed data is available in the fast memory most of the time. Partition and replication is used to co-optimize memory access time and resources.

Abstract: Systems, apparatuses, and methods to identify bus-off and masquerade attacks against electronic control units (ECUs) transmitting on a communication bus from behind a gateway coupled to the communication bus are described. The disclosure further describes systems, apparatuses, and methods to mitigate against bus-off attacks made against an ECU coupled to a communication bus through a gateway. Other embodiments are described and claimed.

Abstract: A system includes a representation generator subsystem configured to execute a user representation model and a task prediction model to generate a user representation for a user. The user representation model receives user event sequence data comprises a sequence of user interactions with the system. The task prediction model is configured to train the user representation model. The user representation includes a vector of a predetermined size that represents the user event sequence data and is generated by applying the trained user representation model to the user event sequence data. A storage requirement of the user representation is less than a storage space requirement of the user event sequence data. The system includes a data store configured for storing the user representation in a user profile associated with the user.

Abstract: A device may receive historical project data identifying experiences and/or work product from previous projects and client data identifying a client with a problem, and may process the historical project data and the client data, with machine learning models, to generate recommendations for the problem and confidence scores for the recommendations. The device may process the recommendations and the confidence scores, with an NLG model, to generate a solution to the problem and content for the solution, and may generate a digitized dynamic client solution to the problem based on the solution and the content. The device may provide the digitized dynamic client solution to a user device, and may receive feedback on the digitized dynamic client solution from the user device. The device may generate a final digitized dynamic client solution based on the feedback, and may perform actions based on the final digitized dynamic client solution.

Abstract: An improved content indexing (CI) system is disclosed herein. For example, the improved CI system may include a distributed architecture of client computing devices, media agents, a single backup and CI database, and a pool of servers. After a file backup occurs, the backup and CI database may include file metadata indices and other information associated with backed up files. Servers in the pool of servers may, in parallel, query the backup and CI database for a list of files assigned to the respective server that have not been content indexed. The servers may then request a media agent to restore the assigned files from secondary storage and provide the restored files to the servers. The servers may then content index the received restored files. Once the content indexing is complete, the servers can send the content index information to the backup and CI database for storage.

Abstract: In one embodiment, a first entry in a first database is modified to include data from a highest-ranked one of one or more available data tables that correspond to the first entry. Each of one or more characters fields of the modified first entry are converted into a respective one or more first-entry tokens, and each of one or more character fields of each of a plurality of second entries in a second database is converted into a respective one or more second-entry tokens. The first-entry tokens are compared to the second-entry tokens, and, in response to the comparison, it is determined whether the first entry matches one of the second entries. In response to determining that the first entry matches one of the second entries, the first entry and the matching second entry are associated with one another in one or both the first and second databases.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for determining a list of recommended items in response to a user query. An embodiment can generate an ordered relevance list of items, and determine an initial reward value based on an array of relevance scores and an array of revenue values corresponding to the ordered relevance list of items, a parameter alpha assigned to the array of relevance scores, and a parameter beta assigned to the array of revenue values. The embodiment can generate a next list of recommended items from an initial list of recommended items, and further calculate a next reward value associated with the next list of recommended items, and determine a list of recommended items in response to the query based on a comparison of the initial reward value and the next reward value.

Abstract: Containers (100), for example boxes, envelopes, mailers, and the like containing one or more deployable cushioning sheets (1500) are disclosed. The deployable cushioning sheets (1500) are connected to at least one wall of the containers (100) and can be deployed to be partially outside of the container for cushioning, for example by wrapping, an object (1600) to be packaged. The deployable cushioning sheets (1500) can contain, for example, paper, cardboard, foam, bubble film, or the like, or they can contain combinations of different types of materials. The containers (100) are useful in methods of packaging, storing, and shipping, among other uses.

Abstract: Systems and techniques are provided for synchronizing DHCP snoop information. In some examples, a method can include, performing, by a first PE device from a plurality of PE devices, DHCP snooping of a first plurality of DHCP messages between a DHCP client and a DHCP server, wherein the plurality of PE devices is part of an ethernet segment for multihoming the DHCP client. In some aspects, the method includes determining, based on snooping the first plurality of DHCP messages, an association between an IP address corresponding to the DHCP client and a MAC address corresponding to the DHCP client. In some examples, the method includes sending, by the first PE device to at least one other PE device from the plurality of PE devices, a first route advertisement that includes the association between the IP address corresponding to the DHCP client and the MAC address corresponding to the DHCP client.

Abstract: Several methods and systems for encoding pictures associated with video data are disclosed. In an embodiment, a method includes determining by a processing module, whether a picture is to be encoded based on at least one of a skip assessment associated with the picture and an encoding status of a pre-selected number of pictures preceding the picture in an encoding sequence. The method further includes encoding by the processing module, a plurality of rows of video data associated with the picture upon determining that the picture is to be encoded, wherein the plurality of rows are encoded based on a pre-selected maximum encoded picture size.

Abstract: A system for generating a centrally located floating three-dimensional image display for a plurality of passengers positioned within a vehicle includes a display adapted to project a multiplexed hologram comprising a plurality of three-dimensional holographic images, and a plurality of beam splitters, one beam splitter individually associated with each one of the plurality of passengers, each beam splitter adapted to receive one of the plurality of holographic images from the display and to reflect the one of the plurality of holographic images to the associated one of the plurality of passengers, wherein, each of the plurality of passengers perceives one of the plurality of three-dimensional holographic images floating at a central location within the vehicle.

Abstract: In one embodiment, a system includes first host machines implementing a public-cloud computing environment, wherein at least one of the first host machines includes a resource manager that provides a public-cloud resource interface through which one or more public-cloud clients interact with one or more virtual machines, and second host machines implementing a private-cloud computing environment, wherein at least one of the second host machines includes one or more private-cloud virtual machines, wherein at least one of the first host machines further includes a private-cloud VM resource provider through which the resource manager interacts with the private-cloud virtual machines, wherein the VM resource provider translates requests to perform virtual machine operations from a public-cloud-resource interface to a private-cloud virtual machine interface, and the private-cloud virtual machines perform the requested virtual machine operations in response to receiving the translated requests from the VM resource

Abstract: A miRNA-mimic based therapeutic particle is disclosed herein. The particles comprise a synthetic miRNA or mimic of miR-187-3p encapsulated in a lipid nanoparticle (LNP) carrier or synthetic miR-193b-5p inhibitor encapsulated in a lipid carrier or their combination encapsulated in a lipid carrier. The lipid nanoparticle carrier is made up of at least four (4) types of lipids, in which the four (4) types of lipids include a) an ionizable cationic lipid selected to be positively charged in a formulation buffer (pH 4), which binds and protects the negatively charged miRNA, and facilitates endosomal escape, and is neutral in a storage buffer, b) a sterol in the structure of the lipid nanoparticle (LNP), c) a structural helper lipid selected to contribute to lipid nanoparticle stability and/or enhances endosomal release, and d) a PEGylated-lipid selected such that it stabilizes the therapeutic particle and protects it from opsonization.

Abstract: The present disclosure describes various aspects of devices and methods that are directed to systems for controlling power systems. A device comprises at least one power source unit having a plurality of physical unit pins, at least one sensor to measure a characteristic of at least one power source unit, and a microcontroller. The microcontroller comprises a memory with at least virtual unit object, at least one virtual sensor object, and a set of software instructions to cause the microcontroller to control the at least power source unit via at least one switch using a three-layer control scheme.

Abstract: Systems and techniques are provided for processing image data. A respective first value enclosed by a convolution kernel in each position of a plurality of positions of the convolution kernel along a row of the image data can be obtained and stored using a respective memory location associated with each position of the plurality of positions. Based on each respective first value, an accumulated value corresponding to a convolution output for each position of the plurality of positions can be updated. A plurality of second values enclosed by the convolution kernel in each position of the plurality of positions can be obtained. The plurality of second values includes a subset of the respective first values and an additional second value. A memory location used to store a first value not included in the plurality of second values can be updated to store the additional second value.

Abstract: Methods, apparatus, and processor-readable storage media for predicting task execution efforts using artificial intelligence techniques are provided herein. An example computer-implemented method includes determining intent information associated with at least a portion of a task by processing data related to the task using at least a first set of artificial intelligence techniques; determining task execution workflow data based at least in part on the intent information associated with at least a portion of the task; predicting one or more efforts associated with executing the task by processing at least a portion of the task execution workflow data using at least a second set of artificial intelligence techniques; and performing one or more automated actions based at least in part on at least one of the one or more predicted efforts associated with executing the task.

Abstract: Systems and methods are provided for distributed video storage and search with edge computing. The method may comprise caching a first portion of data on a first device. The method may further comprise determining, at a second device, whether the first device has the first portion of data. The determining may be based on whether the first piece of data satisfies a specified criterion. The method may further comprise sending the data, or a portion of the data, and/or a representation of the data from the first device to a third device.