21 results about "Network Computing System" patented technology

Malware activity detection for networked computing systems is described. A network session record is provided to a machine learning (ML) model configured to generate an indication of whether the provided network session record evidences malware activity. The network session record indicates network traffic activity in a time period. Responsive to an indication by the ML model, correlation scores are calculated by, for each process session record in a process session record set, calculating a correlation score indicative of a correlation between the provided network session record and the process session record. Each process session record in the process session record set corresponds to a process executed by a computing device in the time period. A determination that a correlation score indicates a corresponding process session record is indicative of the evidenced malware activity is made. Responsive to the determination, a malware activity alert is generated.

The application discloses a kind of diffractive light neural network computing systems and methods for realizing pure optical nonlinearity, belong to photonic computing and artificial intelligence hardware technical field.Its system includes coherent light source, linear light computing unit, nonlinear light activation unit and optical detection unit arranged in order along optical path, wherein linear light computing unit is programmable phase modulation using spatial light modulator, nonlinear light activation unit is programmable binary amplitude modulation using digital micromirror device, and both are directly coupled to form pure optical computing path without photoelectric conversion on optical path.Its method trains neural network through customized loss function, drives nonlinear activation function output to binary convergence, and maps the parameters obtained by training into phase map and binary mask respectively and loads to hardware.The application realizes true all-optical nonlinear computation, with the advantages of high energy efficiency, extremely low delay, compact structure and strong parallel processing capability.

A firing neural network computing system and method for brain-like intelligence and cognitive computing. The system includes a model description module, a parameter database, a configuration description module, a configuration manager, a rule manager, a data manager, a network builder, a network manager, an operation manager, a scheduler, a log manager, an operation monitoring module and a graphical display module. The system provides the function of automatically executing synapse and neuron pruning and genesis according to certain conditions and rules, and provides a variety of flexible trigger conditions for starting up pruning and genesis processes as well as rules for executing the processes, which eliminates the burden of neural network developers needing to write synapse and neuron pruning and genesis programs by themselves, thereby effectively solving several problems in existing brain-like firing neural network computing frameworks.

A system includes a network computing system and a user computing device. The network computing system includes a plurality of software services associated with aircraft operations. The user computing device is operable to connect to the network computing system, over a network, to access data from one or more software services of the plurality of software services. The user computing device is operable to automatically perform one or more pre-flight computing functions for an aircraft based on the data from the one or more software services of the network computing system. At least one of the pre-flight computing functions includes an automatic weight and balance computation for a payload of the aircraft based on the data accessed from the one or more software services of the network computing system.

A computing system is disclosed in which a tensor-network structure mutates autonomously during execution. A structural mutation unit modifies tensor-network topology concur- rently with tensor contraction execution without halting computation or requiring host processor intervention.

Malware activity detection for networked computing systems is described. A network session record is provided to a machine learning (ML) model configured to generate an indication of whether the provided network session record evidences malware activity. The network session record indicates network traffic activity in a time period. Responsive to an indication by the ML model, correlation scores are calculated by, for each process session record in a process session record set, calculating a correlation score indicative of a correlation between the provided network session record and the process session record. Each process session record in the process session record set corresponds to a process executed by a computing device in the time period. A determination that a correlation score indicates a corresponding process session record is indicative of the evidenced malware activity is made. Responsive to the determination, a malware activity alert is generated.

The application discloses a resource high-efficiency utilization in-network computing system, which comprises a task analyzer, a network in computing optimizer and a task placer. The task analyzer is used for automatically summarizing the attributes of each task, including the components and their dependency relationship, so as to obtain the analysis result of the task analyzer output. The network in computing optimizer is used for accepting the network topology and the in-network computing intention and the analysis result of the task analyzer output as the input thereof, taking the number of task copies between switches as the optimization target, and optimizing the task by using the mixed integer programming, so as to obtain the optimal unloading scheme. The task placer is used for enforcing the optimal unloading scheme and ensuring the execution correctness of each task. The in-network computing problem is converted into an optimization problem, so that the task copy copies generated in the existing in-network computing can be reduced, and the resource utilization rate of the in-network computing is increased. Compared with the prior art, the application has faster computing speed and lower overhead. The application is simple and convenient to operate and has strong practicability.

The invention discloses a neural network computing system and device, and relates to the technical field of neural networks, the neural network computing system comprises a digital time conversion module, a digital frequency conversion module and a phase accumulator, and the output end of the digital time conversion module and the output end of the digital frequency conversion module are connected with the input end of the phase accumulator; the digital time conversion module is used for compressing and converting sparse data in the received input data to obtain a corresponding time pulse signal and outputting the time pulse signal to the phase accumulator; the digital frequency conversion module is used for converting the received weight data into corresponding frequency signals and outputting the frequency signals to the phase accumulator; and the phase accumulator is used for carrying out multiply-accumulate calculation on the time pulse signal and the frequency signal based on the phase and outputting a calculation result. According to the method, the technical problems of high resource overhead and high energy consumption of a neural network calculation strategy aiming at sparse data at present are solved.

In conventional networked computing systems, stream-processing software is utilized to handle real-time data feeds. For example, a stream-processing software may sort incoming data feeds into data logs, which are then consumed by various application servers within the computing system. However, this consumption process is susceptible to failure as application servers often fail to consume a full payload of data comprising a particular message. As such, a need exists for a streaming architecture with improved fault tolerance. The system provided herein solves the problem via a reliable streaming process, which regularly generates partitions within a series of data lakes. As streaming data is received, the payloads are synced into the different data lakes based on metadata associated with the streaming data. The system can then perform individual quality checks on each partition of data, enabling the system to continuously monitor data quality over time.

A computer implemented method manages a topology for a network computing system. A number of processor units determines tasks for a workload to be performed by computing devices in the network computing system. The number of processor units creates the topology for performing the tasks in the workload. The topology comprises nodes representing the tasks assigned to the computing devices. The nodes are connected to each other based on an execution order for the tasks, and the topology has alternative paths for performing the tasks for the workload. The number of processor units determines energy consumption for the alternative paths in the topology. The number of processor units selects a particular path from the alternative paths for performing the tasks for the workload based on the energy consumption for the alternative paths.

A method of a computing system of a communication network for providing video to a portable display device includes establishing a connection with a receiving system comprising the portable display device, obtaining compressed video data and decoding the compressed video data, receiving direction information from the receiving system, the direction information related to a direction of the portable display device, identifying a respective location of each macroblock of a plurality of macroblocks of the decoded video data, determining a respective distance between a location determined based on a field of view identified utilizing the direction information, and the respective location of each macroblock of the decoded video data, encoding the decoded video data utilizing a respective quantization parameter identified based on the respective distance determined for each macroblock to provide re-encoded video data, and sending the re-encoded video data to the receiving system for display on the portable display device.

A network computing system receives, over one or more networks, a service request from a requester device. The network computer system predicts a processing time interval, where the processing time extends from a time when the service request is received until a time when a service provider that is selected by the network computer system accepts an invitation to fulfill the service request. The network computer system transmits content data to the requester device, where the content data causes the requester device to display content that is based on the predicted processing time interval.

The application discloses a kind of Sub-GHz-based airborne wireless sensor network node system and matching network calculation method, including sensor sensitive unit module, single-chip microcomputer control module, wireless communication module and power supply module;With the compatibility between the original equipment of Sub-GHz-based wireless communication module design ensures airborne, optimize impedance matching circuit makes transmission power reach 20dBM, center frequency 915MHz, bandwidth 530MHz, at the suppression of secondary harmonic frequency and third harmonic frequency reaches-40dB below, wireless module maximum transmission rate can reach 2Mbps, and through electromagnetic compatibility test ensures the anti-interference of wireless communication module.The application has strong anti-interference ability, signal attenuation is slow, and low power consumption is suitable for solving cabin complex environment.

Systems, methods, and devices for data ingestion, database management, and data security. A method includes determining that a receiver device is offline and cannot communicate with a server. The method includes providing a plurality of data packets to a data packet queue and receiving an update to an existing data packet stored within the data packet queue. The method further includes minimizing a total memory requirement for the data packet queue by replacing the existing data packet with the update and in response to determining the receiver device is now online, communicating the data packet queue to the receiver device over a network.

Computing systems methods, and non-transitory storage media are provided for obtaining images, extracting layers from each of the images, extracting segments from each of the layers, generating a compressed version of the segments by storing a single copy of each segment and metadata to reconstruct the layers from the segments and the images from the layers, and simulating a reconstruction of the image from the compressed version.

The application discloses an automatic driving model confrontation training method based on double time scales and Nash equilibrium, and belongs to the technical field of automatic driving. The method comprises the following steps: establishing a strategy pool and setting an asymmetric learning rate; performing closed-loop simulation and experience collection; in the inner loop, the number of update steps of a test network is greater than that of a host vehicle network; constructing a metagame payoff matrix and solving a mixed strategy Nash equilibrium; taking the Nash equilibrium mixed model as a teacher model to perform knowledge distillation to a single network; calculating three health indicators of system availability, win rate climbing rate and strategy convergence stability and adaptively adjusting hyperparameters. The application eliminates catastrophic forgetting through a double time scale asymmetric update mechanism, avoids mode collapse through Nash equilibrium solving, realizes vehicle-mounted low-power deployment through knowledge distillation, realizes automatic adaptive training through health indicator monitoring, and significantly improves the stability and robustness of automatic driving model confrontation training.

The application discloses a kind of multi-layer nonlinear optical neural network computing systems based on linear signal re-encoding, belong to optical neural network computing technical field, including: control circuit and optical processing chip system;The control circuit is used to realize the loading of input signal and the reading of output result;The optical processing chip system includes light source, integrated optical computing chip and detector, and the light carrier emitted by light source generates input optical signal after being modulated by input signal, and after being processed by multi-layer nonlinear activation in linear calculation area, linear re-encoding area and high-speed modulation area on integrated optical computing chip, output result is read by detector and returned to control circuit.The application utilizes pure linear device to realize all-optical nonlinear network, cooperates analog activation function by re-encoding and modulation, needs only single physical layer to construct depth architecture without photoelectric conversion, with the advantages of all-optical processing, CMOS compatible, high speed and low power consumption.

The invention discloses a tensor processing method used in a neural network computing system. The tensor processing method comprises the following steps: creating a computational graph descriptor for each tensor in the neural network computing system; when a source tensor needs to be copied or a band is copied to a target tensor, acquiring a read computational graph of the source tensor from a computational graph descriptor of the source tensor, and acquiring a write computational graph of the target tensor from a computational graph descriptor of the target tensor; splicing the output end of the reading calculation graph of the source tensor and the input end of the writing calculation graph of the target tensor to form a complete carrying calculation graph, and performing optimization operation on the carrying calculation graph; and generating execution code for the target hardware according to the optimized transport computational graph.

A system includes a network computing system and a user computing device. The network computing system includes a plurality of software services associated with aircraft operations. The user computing device is operable to connect to the network computing system, over a network, to access data from one or more software services of the plurality of software services. The user computing device is operable to automatically perform one or more pre-flight computing functions for an aircraft based on the data from the one or more software services of the network computing system. At least one of the pre-flight computing functions includes an automatic weight and balance computation for a payload of the aircraft based on the data accessed from the one or more software services of the network computing system.

An information sharing system with a display screen and a network computing system is disclosed. The display screen may be capable of either being manually moved by a person or move automatically under its own power. The apparatus also includes a power system, a communication subsystem, a memory and a computing device. The communication subsystem communicates with the computing device, the network computing system, and the memory. The network computing system has an end-user display screen. The network computing system is configured to receive media and media secondary information, associate the media and media secondary information to a unique identifier, and show the media and the unique identifier on the display screen. The networking computing system is further configured to receive the unique identifier and an end-user account, and then show the media secondary information on the end-user display screen.

In some embodiments, a server system may re-allocate a resource to a node from a shared pool of available resources and reschedule the added allocation. Some embodiments may use such operations to avoid triggering application-terminating thresholds and increase the efficiency and consistency of a networked computing system. In response to obtaining a result indicating whether an action will cause a resource allocation level of a node to fail to satisfy a threshold, some embodiments may delay the action by a buffer duration and allocate a set of resources from a resource pool to the node. As such, in some embodiments, the set of resources may be available to the node before an end of the buffer duration. Thus, at the end of the buffer duration or after the buffer duration, some embodiments may execute the action without triggering a failure condition related to the node.