Abstract: A computer-implemented method for performing natural language-based data integration includes causing execution of a data integration application on a remote device via a network and causing surfacing of a GUI corresponding to the data integration application on a display of the remote device. The method includes receiving, via the GUI, a natural language input representing a data integration task, generating, via an LLM, a set of ordered activities corresponding to the data integration task represented by the natural language input, and selecting, via the LLM, one or more APIs for performing each activity within the set of ordered activities. The method also includes generating a data pipeline based on the set of ordered activities and the API(s) for performing each activity, as well as back-translating the data pipeline to a desired data format for execution by the data integration application.

Abstract: As understood herein, not all objects in a computer game are of equal priority. Given this recognition, present principles apply a range of spatial and temporal resolutions to various objects in the same game depending on the priority of the objects in the game.

Abstract: Techniques are described for determining if chat between two computer gamers requiring translation is related to the game or is simple social chat. Depending on the type of chat, a machine learning (ML) model is selected to predict completion of the sentence being input by the speaker to reduce latency in translating and transmitting the chat.

Abstract: Techniques are described for reducing latency in networked gaming by reducing I-frame sizes (which also results in automatically increasing P-frame sizes) to reduce the overall amount of video being transmitted. The reduced size of the I-frames is compensated for by increasing the size of other frames using a low pass filter (LPF) such as a Gaussian filter which reduces sharpness that the decoder can try to recover, or by use of lower resolution. The I-frame can be reduced by rotating it or flipping/mirroring it to produce the smaller coded frame, sending a flag to signal the orientation.

Abstract: Techniques are described for over-training a ML model on multiple gameplay videos of individual scenes of a computer game to better configure the model to reconstruct or enhance portions of the computer game at a receiver as the computer game is received over a streamlining network. Reconstruction of individual missing slices of a frame is contemplated such that a frame missing a slice need not be entirely discarded.

Abstract: Techniques are described for translating computer game chat to a different language for a partner gamer. If the chat is determined to be related to a game being played as opposed to social chat, a glossary may be used tailored to that game to first determine if a spoken term is in the glossary. If it is, a lookup of an equivalent term from the glossary can be used instead of passing the term through a translation engine, increasing accuracy and saving translation time.

Abstract: Multiple (e.g., two) hosts access a single memory channel (and/or device) via a memory controller. The single memory channel/device can support at most one access at a time. To reduce contention between the multiple hosts, the memory controller comprises multiple (e.g., two), independent, host ports. Each host port is associated with a write buffer(s) in the memory controller that stores write data at least until the memory controller writes the data to the memory channel. Data stored in a write buffer may be used to respond to memory access commands (e.g., reads or writes) on the ports without accessing the memory channel. In this manner, the hosts do not directly contend with each other for the single memory channel or the memory controller.

Abstract: A computer-implemented method for performing natural language-based data integration includes causing execution of a data integration application on a remote device via a network and causing surfacing of a GUI corresponding to the data integration application on a display of the remote device. The method includes receiving, via the GUI, a natural language input representing a data integration task, generating, via an LLM, a set of ordered activities corresponding to the data integration task represented by the natural language input, and selecting, via the LLM, one or more APIs for performing each activity within the set of ordered activities. The method also includes generating a data pipeline based on the set of ordered activities and the API(s) for performing each activity, as well as back-translating the data pipeline to a desired data format for execution by the data integration application.

Abstract: A DMA system includes two or more DMA engines that facilitate transfers of data through a shared memory. The DMA engines may operate independently of each other and with different throughputs. A data flow control module controls data flow through the shared memory by tracking status information of data blocks in the shared memory. The data flow control module updates the status information in response to read and write operations to indicate whether each block includes valid data that has not yet been read or if the block has been read and is available for writing. The data flow control module shares the status information with the DMA engines via a side-channel interface to enable the DMA engines to determine which block to write to or read from.

Abstract: Multiple (e.g., two) hosts access a single memory channel (and/or device) via a memory controller. The single memory channel/device can support at most one access at a time. To reduce contention between the multiple hosts, the memory controller comprises multiple (e.g., two), independent, host ports. Each host port is associated with a write buffer(s) in the memory controller that stores write data at least until the memory controller writes the data to the memory channel. Data stored in a write buffer may be used to respond to memory access commands (e.g., reads or writes) on the ports without accessing the memory channel. In this manner, the hosts do not directly contend with each other for the single memory channel or the memory controller.

Abstract: A method, system, and computer readable medium for improved decoding CABAC encoded media are described. A decoded binary string is output using two or more previously decoded elements at an end state for a decoding loop and on a first processing thread. A syntax of the decoded binary string is parsed on a second processing thread and a decoded symbol from the parsed syntax is generated on the second processing thread.

Abstract: A method, system, and computer readable medium for improved decoding CABAC encoded media are described. A decoding loop includes decoding an encoded binary element from a sequence of encoded binary elements to generate a decoded binary element using a context probability. A next context probability for a next encoded binary element in the sequence is determined from the decoded binary element and the next context probability for decoding the next encoded binary element is provided to the decoding loop for a next iteration.

Abstract: The present invention relates to the field of fuel cell stacks and stack tower or module, in particular to a sealing structure for stack tower and a stack tower. The sealing structure comprises a first component, a second component and a mica spacer, the first component and the second component are opposite to each other, the mica spacer is disposed between the first component and the second component, sealing part is arranged between the mica spacer and at least one of the first component and the second component, and the sealing part comprises a glass ceramic layer and an outer circumferential ceramic cement ring surrounding the glass cement layer; the sealing structure for stack tower has excellent sealing performance and service durability for a fuel cell system.

Abstract: An image of a computer game player wearing a headset that occludes part of the face is input to a trained machine learning (ML) model, which outputs in response a full-face image that is not occluded for use in, e.g., social network settings related to the game.

Abstract: A method, computer system, and computer program product are provided for integrating access to multiple cloud storage sources from different applications. A storage account is joined to a local domain. The storage account is registered with a directory service in the local domain. A first service accesses the storage account by using a domain credential for accessing the local domain. A second service accesses the storage account by using a directory service credential to the directory service. The first service and the second service may concurrently access the storage account.

Abstract: A multi-assistant speech-processing system that centrally determines multiple execution plans to respond to a user input. A central component determines whether a particular input should be processed using a requested assistant or a different assistant or should be terminated. Assistant handoff may be determined based on system policies as well as user input-specific data. A ranked list of execution options may be supplemented by augmented data corresponding to messages to a user. The system may attempt to execute plans in the ranked order until a plan succeeds.

Abstract: A computing system includes a host device and a root of trust (RoT) device for performing batch encryption and decryption operations facilitated by a direct memory access (DMA) engine. The host device generates a command table for batch processing of a set of address tables that each describe a set of data blocks of a file to be encrypted or decrypted. The DMA engine facilitates a DMA transfer of the command table from the host memory to an RoT memory of the RoT device. The RoT device then performs batch processing of the address tables referenced in the command table. For each address table, the DMA engine copies a set of data blocks from the host memory to the RoT memory, a cryptographic engine encrypts or decrypts the data blocks, and the DMA engine copies the transformed data blocks back to the host memory.

Abstract: Application state data from a main memory may be compressed and the compressed data may be written to a first location in a mass storage. Updated application state data is generated, and the updated application state data is compressed from the main memory. The updated application state data is then written to a second location in the mass storage. Processing may then be paused on the application state data and updated application state data. The compressed application state data and compressed updated application state data stored in the mass storage is scanned and information corresponding to compressed application state data and updated compressed application state data stored in the mass storage is displayed using information from the scanned compressed application state data and compressed updated application state data.

Abstract: Technologies for detecting and classifying errors detected in pipelined hardware are described. One device includes a hardware pipeline with a set of pipeline stages. Error detection logic can detect an error in the hardware pipeline, and control logic can classify the error in one of the multiple categories based on a type of the error, a position of the first data in a data stream that triggered the error, and a position of a pipeline stage in which the error is detected. The control logic can perform an error-response action based on the error classification of the error.