95 results about "Algorithm configuration" patented technology

The invention belongs to the technical field of artificial intelligence, is applied to the field of financial science and technology and the field of medical health, and discloses a rule engine configuration method, device and equipment based on machine learning and a storage medium. The method comprises the steps that historical calculation data and performance indexes in historical calculation tasks are collected, and a training data set is constructed; a machine learning model is trained, and algorithm parameters of the rule engine are dynamically optimized through the training data set; configuring an expression and a user-defined function of a rule engine, and realizing analysis and execution of scripts through JEXL (JavaScript Exchange Library); establishing a dynamic mapping relationship between the algorithm configuration library and the business variables, mapping algorithm variable names and business variable names, and generating configurable algorithm rules; and analyzing task load and resource requirements based on a pre-trained machine learning model, distributing system resources, and repairing abnormal behaviors in a calculation process. According to the invention, the problems of low processing efficiency, high maintenance cost, unreasonable resource allocation and poor expandability in the prior art are solved.

Methods for a machine-learning network that provide efficient, scalable, and granular analyses during validation of a machine learning model are disclosed. Validation of models depends upon many factors, including the real-world application of the model, the type of model being trained, and the types of data samples it is being trained on. In order to provide relevant edge case information to users that pertains to their specific model, data slice finding techniques may be used to identify subsets of the dataset that are particularly problematic. By limiting a length of the slice description that the algorithm searches and by configuring the algorithm to target specific types of errors, users are provided with a more granular analysis that then allows them to determine how or if they need to retrain the model.

The invention discloses a reinforcement learning aided evolutionary algorithm for parameter optimization in algorithm configuration, which comprises the following steps of: performing parameterized definition on four core operators, coding all adjustable parameters, and constructing a unified parameter configuration space; then randomly initializing a configuration population, and inputting the configuration population into an optimization framework assisted by reinforcement learning; the framework is internally provided with an intelligent agent and is executed in each generation of iteration: operating an evolutionary algorithm according to parameter configuration and evaluating performance; the intelligent agent adaptively selects a predefined alignment strategy based on a population state, and guides mutation operation to generate a filial generation; performing environment selection and population updating based on performance evaluation; and finally, calculating rewards according to population performance, and training and updating the agent decision model. And after iteration to a termination condition, outputting configuration with optimal performance and a corresponding algorithm thereof. According to the method, an evolutionary algorithm search space is constructed through comprehensive parameterization, a reinforcement learning agent is introduced to adaptively process variable-length coding, and a double-layer optimization framework is adopted to efficiently search optimal algorithm configuration.

The invention discloses an industrial image anomaly detection method based on a multi-agent arrangement heterogeneous algorithm. The method comprises the following steps: S1, constructing an algorithm component library for shielding data representation differences among different architecture algorithms; s2, constructing a multi-modal large model system based on a multi-agent architecture; the multi-agent architecture comprises the following steps: receiving an image by using a visual expert agent, and outputting structured physical metadata and an unstructured visual suggestion text; the planner agent receives the physical metadata and the visual suggestion text, and generates an algorithm configuration tree based on an algorithm component library; the optimizer agent optimizes the algorithm configuration tree in the closed-loop feedback stage; s3, using the algorithm configuration tree to construct a reference model representing good product distribution according to the normal training sample images; s4, performing anomaly detection on the to-be-detected sample image by using the algorithm configuration tree to generate an anomaly judgment result; and S5, optimizing the algorithm configuration tree.

The invention relates to the technical field of adaptive image processing, in particular to a satellite remote sensing image processing adaptive optimization method based on resource constraint, and aims to improve satellite processing efficiency. The method comprises the following steps: acquiring task information which is uploaded by a ground system and contains target product information; determining a first processing flow according to the task information; the first processing flow comprises a plurality of steps of producing a target product, each step corresponds to one algorithm sub-category, and each algorithm sub-category has a selectable identifier or a necessary identifier; assigning a corresponding algorithm module for each algorithm sub-category in the first processing flow according to an algorithm configuration table so as to obtain a second processing flow; the second processing flow is optimized according to the selectable identifier and the necessary identifier, so that the estimated value of each constraint term of the optimized processing flow is smaller than or equal to the corresponding upper limit value; and executing the optimized processing flow to generate a target product. According to the method provided by the invention, the on-satellite processing efficiency is improved, and the information extraction precision is also considered.

The invention discloses an automatic algorithm configuration method based on deep reinforcement learning in evolutionary computation. The method comprises the following steps: determining a problem set; obtaining initial data for constructing a Markov decision process; defining a state vector of each individual in a time step in an evolutionary computation algorithm by using an individual state vector according to the initial data; constructing feature vectors of population features, individual features and development features according to the initial data; defining a continuous action space according to the initial data, and jointly controlling the selection of all individuals on hyper-parameters; based on whether a better solution is found or not, a reward mechanism is set, and maximization of expected benefits is achieved; and the deep reinforcement learning agent uses an attention mechanism to control dynamic hyper-parameters of the algorithm, so that a Markov decision process is realized, and automatic algorithm configuration is completed. According to the method, direct mapping from high-dimensional sensing information to continuous action space output is realized, configuration of individual exploration and development tradeoff is dynamically adjusted, and a balance mechanism of exploration and development is optimized.

The invention relates to the technical field of semiconductor manufacturing, in particular to a semiconductor scheduling algorithm model configuration system, method and equipment and a medium. The system comprises a data configuration module, an algorithm configuration module, a parameter analysis module, a task scheduling module, a data acquisition module, a model execution module and a result processing module. The data configuration module receives data model configuration parameters input by a user; the algorithm configuration module receives algorithm model configuration parameters. The parameter analysis module analyzes parameters and generates a data preprocessing rule and an algorithm model instance, the task scheduling module creates a scheduling task and triggers a process, the data acquisition module acquires original data from a source data table and preprocesses the original data, the model execution module executes a scheduling algorithm based on the algorithm instance, and the result processing module outputs standard data. According to the invention, decoupling of the data and the algorithm is realized, and the adaptability and expandability of the system are improved.

The application discloses a cryptographic algorithm debugging data hierarchical grouping compression method for a reconfigurable cryptographic chip. In the debugging work of the reconfigurable cryptographic chip, the debugging data is divided into reconfigurable cryptographic algorithm configuration information and operation intermediate data. The debugging data is processed by hierarchical grouping, the debugging data of each level module is divided into one or more groups, then the debugging data is filled or spliced, and then the reconfigurable debugging data is compressed by using a Huffman coding mode. After the hierarchical grouping compression method of the debugging data is used, the transmission performance of the debugging data can be effectively improved while the storage resource of the debugging data is reduced.

The invention discloses a configuration method and device for monitoring equipment in multiple scenes, and relates to the field of artificial intelligence and big data, and the method comprises the steps: obtaining K business scenes of target monitoring equipment, and determining a monitoring demand for each business scene; determining a to-be-configured target algorithm based on the monitoring requirement of the target monitoring equipment in each business scene, and configuring algorithm parameters and an algorithm analysis mode for the target algorithm; determining an alarm triggering rule based on the algorithm parameters; and generating an algorithm analysis task based on the business scene, the monitoring demand, the target algorithm, the algorithm parameter, the algorithm analysis mode and the alarm triggering rule, and issuing the algorithm analysis task to the target monitoring device. According to the method and the device, the technical problem of low reuse rate of a mode of performing single algorithm configuration on the monitoring equipment in related technologies is solved.

This application provides a method for configuring video stream algorithms, comprising: extracting visual features, spatiotemporal features, and semantic features of the input video stream; fusing these features to generate a first scene feature of the video stream; and generating a configuration algorithm for the video stream based on the first scene feature and a pre-constructed knowledge graph representing the relationship between scene features and algorithms. The technical solution of this application can effectively solve the problems of high configuration cost, poor scene adaptability, knowledge transfer gaps, and low configuration efficiency of video stream algorithms in smart city scenarios in existing technologies.

There is provided a computer implemented method comprising obtaining image data representing a platform, processing the image data using an algorithm, the algorithm configured to determine a property of the platform, obtaining, from the algorithm, data indicating a property of the platform and determining, based on the data indicating the property of the platform, a location to which the platform is to be transported.

The invention provides a performance analysis method and device for source address verification, equipment and a storage medium, and relates to the technical field of network security. The method comprises the following steps: receiving a source address verification processing program written through a description language; loading an algorithm configuration file; wherein the algorithm configuration file at least comprises a performance model configured for a plurality of preset algorithms, and the performance model is used for calculating the consumed time length of the operation statement based on the operand length of each operation statement; based on the source address verification processing program and the performance model, calculating the consumed time length of each operation statement; and generating a performance analysis report of the source address verification processing program according to the time-consuming duration of all the operation statements in the source address verification processing program. According to the embodiment of the invention, automatic and theoretical performance analysis can be carried out on the source address verification processing flow based on the formalized description language and the predefined algorithm configuration file under the condition of separating from the dependence of specific hardware.

The invention discloses an encryption device and system, and the device comprises a global encryptor which is used for obtaining algorithm configuration data which is determined through the joint negotiation of a producer and a consumer based on an agreed target cryptographic algorithm, and constructing a target encryptor instance which supports the target cryptographic algorithm; the encoder conforms to serializer interface specifications of a client library of a producer and is used for encoding first data generated by the producer into a first plaintext byte array; the target encryptor instance is called to encrypt the first plaintext byte array into a first ciphertext byte array through a target cryptographic algorithm, so that a producer stores the first ciphertext byte array to a distributed stream processing platform; and the decoder conforms to a deserializer interface specification of a client library of the consumer and is used for calling the target encryptor instance to decrypt a second ciphertext byte array obtained by the consumer from the distributed stream processing platform into a second plaintext byte array through a target cryptographic algorithm and decode the second plaintext byte array into second data.

The invention relates to a supervision algorithm calling method and device, computer equipment, a medium and a product. The method comprises the following steps: receiving field operation information of a current task sent by terminal power equipment and algorithm configuration information sent by a management platform; determining a target algorithm corresponding to the field operation information; determining a target server corresponding to the field operation information according to the algorithm configuration information and a target algorithm; the algorithm configuration information comprises a corresponding relation between algorithms and servers, and one server corresponds to at least one algorithm; sending the field operation information to a target server; receiving a processing result sent by the target server; the processing result is obtained by processing the field operation information by the target server based on the target algorithm. By adopting the method, real-time and accurate supervision of the field operation process can be realized, and the strict requirements of modern industry on operation safety, high efficiency and compliance are met.

The disclosure provides methods and apparatus for training and executing a reinforcement learning, RL, algorithm configured with multiple reward functions to perform a task in a communication network (1000) The method (100, 500) comprising, the first network node (1100) transmitting (103, 503) a first message, to the second network node (1200), comprising parameters of the RL algorithm. The first network node (1100) receiving (104, 507) a second message, from the second network node (1200), comprising one or more observations. The first network node (1100) updating (105, 508) the parameters of the RL algorithm based on the received one or more observation. The first network node (1100) performing (106, 509) the transmitting, the receiving and the updating iteratively until a stopping criterion exceeds a threshold and storing (107, 510) the updated parameters of the RL algorithm.

The invention provides an old city pattern recognition and updating method and system based on multi-source data fusion, and the method comprises the steps: collecting a multi-source data set, analyzing n-hour activity cycle characteristics, and generating a time period thermodynamic diagram; identifying a community structure by using a graph neural network, outputting social topology, and quantifying a texture evolution track by using a dynamic attitude model; quantizing the traditional spatial pattern by using a landscape sight line algorithm, and outputting a dynamic pattern recognition result; and configuring aging facilities by using a population adaptation algorithm, optimizing industrial layout by using a location entropy algorithm, and generating a micro-update scheme. According to the method, a space-time-society-ecology four-dimensional dynamic pattern recognition system is constructed; spatial pattern quantization and weight adjustment are realized by innovatively using a landscape sight line algorithm, and the traditional limitation is broken through; and finally, a tensor decomposition algorithm is used for fusing the configuration and optimization scheme, a micro-update scheme is generated, an update strategy is formed through iterative optimization, complex requirements are accurately responded, and the old city pattern recognition scientificity and the update scheme implementation efficiency are improved.

The invention is suitable for the field of data processing, and discloses a method for compressing data between a base station and a core network, terminal equipment and a storage medium. The data compression method is applied to a data compression system, the data compression system comprises a base station side compression / decompression module and a core network side compression / decompression module, and the data compression method between the base station and the core network comprises the following steps: receiving compression configuration information sent by the core network, the compression configuration information comprising a service type; obtaining compression capability information of the base station, and generating a compression algorithm configuration parameter set according to the compression configuration information and the compression capability information; and when target data transmitted between the base station and the core network is detected, performing compression operation on the target data according to the compression algorithm configuration parameter set. According to the invention, intelligent compression of transmission data between the base station and the core network is realized, the data volume is effectively reduced, the method is suitable for a satellite link with a limited transmission rate, and the problem of waste of satellite resources is relieved.

The invention discloses a VPN (Virtual Private Network) system encryption algorithm management and control method and device, relates to the technical field of network security, and mainly aims to overcome the limitation of fixed configuration of a VPN system encryption algorithm and realize dynamic configuration of a VPN tunnel encryption algorithm. According to the main technical scheme, the method comprises the steps that the VPN system configures an algorithm calling engine corresponding to at least one algorithm library, and the algorithm calling engine executes encryption operation by calling encryption algorithms in the corresponding algorithm library; if it is determined that the VPN system establishes the tunnel, algorithm description information corresponding to the configured algorithm library is sent to both communication parties of the tunnel, and the algorithm description information is used for describing an encryption algorithm in the algorithm library; and if it is determined that the two communication parties negotiate a target algorithm library required to be used by the tunnel based on the algorithm description information, calling an encryption algorithm in the target algorithm library through an algorithm calling engine corresponding to the target algorithm library to encrypt data transmitted in the tunnel.

An apparatus includes a memory and a processor. The memory stores a machine learning algorithm configured to classify telemetry data into a set of categories. The processor implements a communication synchronization scheme to receive a first set of telemetry data associated with a first user and a second set of telemetry data associated with a second user. The processor applies the machine learning algorithm to each of the first and second sets of telemetry data, to classify the data. The processor transmits, to a server, training data that includes at least the classified data or a set of parameters derived from the classified data. The server uses the training data to refine a reinforcement learning algorithm that is configured to generate a recommendation of computational resources to provision to a new user.

The invention discloses a resource optimization method and system for decision making of a main processor and scheduling in an audio processor. The method comprises the following steps: acquiring processor load information when the audio processor performs audio processing on multiple audio input channels and channel feature information of each audio input channel; when the processor load information meets a preset load condition, the main processor determines an adjustment result of audio processing configuration according to a preset resource optimization rule based on the processor load information and the channel feature information; the main processor generates and sends a control instruction to the audio processor according to the adjustment result; and the audio processor updates the sound mixing routing relationship in the internal scheduling according to the control instruction, and carries out switching or intensity adjustment on the processing link of each channel according to the processing algorithm configuration. According to the technical scheme, the continuity and the stability of conference audio processing are improved on the premise of not changing a hardware structure.

The invention discloses an MZI-based space division multiplexing convolution parallel computing method, which comprises the following steps of: configuring an MZI array through an in-situ training algorithm, and enabling the MZI array to respectively present target transmission characteristics corresponding to N preset convolution kernels for incoherent light with N preset wavelengths in a single control state. When calculation is carried out, an input vector to be processed is coded into incident power of light of N wavelengths and injected into the array, the superposition value of the light power of each wavelength is directly detected at an output port of the array, and the superposition value is an N-path parallel convolution result. The system of the invention correspondingly includes an incident layer, a multi-wavelength operating MZI array, and a summing layer. According to the method, the calculation parallelism degree can be improved by N times without increasing hardware overhead, the unitary matrix limitation of a traditional MZI array is broken through, the calculation density and the hardware utilization efficiency are remarkably improved, and the method has wide application prospects in the field of photon calculation.

The embodiment of the invention relates to an information aggregation processing method and device, equipment and a medium, and the method is applied to an aggregation server side and comprises the steps that an event corresponding to an abnormal file is obtained, the event is obtained by analyzing the abnormal file through an analysis server side, and the analysis server side and the aggregation server side are independent of each other and communicate with each other; an aggregation algorithm corresponding to the event type of the event is obtained, and the aggregation algorithm is dynamically configured through an aggregation engine; and performing aggregation processing on the events based on an aggregation algorithm. According to the method, the aggregation server and the analysis server which are mutually independent and communicate with each other are arranged, analysis and aggregation processing are completely separated, the influence of analysis on aggregation during large flow is avoided, the efficiency of event aggregation is effectively improved, an independent aggregation engine is introduced for dynamic configuration of an aggregation algorithm, code modification and release are not needed, and the efficiency of event aggregation is improved. The algorithm configuration cost is greatly reduced, and the algorithm configuration efficiency and flexibility are improved.

The invention relates to a signature algorithm updating method and device, a storage medium and electronic equipment, and the method comprises the steps: initiating a signature algorithm updating request which carries algorithm configuration of a new signature algorithm, a latest signature algorithm library hash value and target block chain node information; receiving update responses of other block chain nodes to the signature algorithm update request; after determining to agree to update according to the received update response, issuing the signature agree to update and the algorithm configuration of the new signature algorithm to the smart contract of the block chain system; the block chain system triggers an update configuration event according to presetting, so that each block chain node in the block chain system updates a local signature algorithm library according to the update configuration event; and each block chain node obtains the time for activating the new signature algorithm in the updated local signature algorithm library according to the smart contract of the block chain system. The transaction efficiency of the block chain system can be improved.

The application discloses a general neural network creation method, device, platform, medium and equipment, and belongs to the field of deep learning. The method comprises the following steps: processing parameters according to custom functions of a user, creating an algorithm configuration file, and integrating the algorithm configuration file into a model package; setting data processing logic of the algorithm configuration file by using the processing parameters in one or more algorithm configuration files, obtaining an algorithm inference library matched with the model package; creating a model package loading component according to a path of the model package; running a specified model in the model package through a model loading interface by using the model package loading component, and setting a default running device; running the specified model in the model package through the algorithm inference library in the default running device, processing user-defined data to obtain network processing data, and creating a general neural network. The efficiency of the general neural network in running the algorithm inference library to obtain a conversion result is higher, the effect is better, and the time for actual deployment of a vehicle is saved.

The invention discloses a model training data processing method and system based on an improved YOLO algorithm. The method relates to the technical field of YOLO algorithm data processing, and comprises the steps of data conversion and original data configuration, PC end prediction recognition model generation, embedded end prediction recognition model generation and embedded end prediction recognition model updating. The method comprises the following steps: extracting a predefined data set to identify and count original parameters configured by a YOLO algorithm; configuring initial parameters according to a YOLO algorithm, and inputting the initial parameters into a training script for data training; converting the PC end prediction identification model into an embedded application environment model; and configuring initial parameters according to the updated YOLO algorithm to reconfigure the embedded end prediction identification model. The effect of improving the application extension adaptability of the YOLO model on the embedded platform is achieved, and the problem that in the prior art, the application extension adaptability of model training data processing of a YOLO algorithm on the embedded platform is insufficient is solved.

The application discloses an intersatellite link satellite data transmission task scheduling method and system and a storage medium, and comprises the following steps: acquiring a satellite set, a task target set and a ground station set; determining a network connection plan and a data transmission plan in each time slot; acquiring a task constraint condition, establishing an intersatellite link satellite data transmission task scheduling model, and taking the maximum complete task data value transmitted to the ground station after T time slots as the scheduling model; constructing a parallel algorithm group, acquiring an optimal algorithm and corresponding parameter configuration, and matching the optimal algorithm with a scene instance. The method constructs an integer programming model of the intersatellite link satellite data transmission task model, constructs a parallel algorithm group, searches for an optimal algorithm configuration in a machine learning mode, and overcomes the contradiction between the cost and quality of huge computing resources required by parallelism.

The invention discloses an industrial detection method based on a cloud and edge collaborative architecture, which relates to the technical field of cloud platform services and mainly comprises the following steps: creating industrial detection items according to a hierarchical structure on a cloud platform; under the task, defining or quoting at least one operator encapsulated with a single data processing function; under the task, at least one workflow is created, and each workflow defines a specific algorithm execution process for the detection scene through one or more operators selected by combining; on the cloud platform, packaging complete algorithm configuration information which is associated with a task and comprises at least one workflow and an operator combined by the workflow to generate a standardized deployment pipeline file; and receiving and analyzing the deployment pipeline file on the client device of the edge end, connecting the industrial detection device, carrying out industrial detection on the detection data collected by the industrial detection device, and outputting a detection result. According to the method, the problem of repeated development in a traditional scheme is solved, and the algorithm construction efficiency is improved.

The invention discloses a simulation setting system and method based on a distribution network digitization system, and belongs to the technical field of digitization simulation, and the simulation setting system comprises a digitization base module which is used for constructing a power distribution network digital twinborn body; the topology construction module is used for dynamically constructing a simulation subnet topology based on the target node of the simulation task and the electrical sensitivity; the model configuration module is used for configuring a simulation mathematical model for the power grid element model according to the parameters of the power grid element model; the algorithm configuration module configures calculation logic for the simulation task according to probability distribution characteristics of input variables in the simulation request; the task packaging module integrates simulation subnet topology, a simulation mathematical model, computational logic and initial parameters, and generates a simulation task package used for being called by an external simulation system; according to the method, the full-process automation and intelligentization of simulation setting are realized, and the simulation setting efficiency and the model fidelity are greatly improved.

Principal component analysis (PCA) accelerates and increases accuracy of genetic algorithms. In an embodiment, a computer generates many original chromosomes. Each original chromosome contains a sequence of original values. Each position in the sequences in the original chromosomes corresponds to only one respective distinct parameter in a set of parameters to be optimized. Based on the original chromosomes, many virtual chromosomes are generated. Each virtual chromosome contains a sequence of numeric values. Positions in the sequences in the virtual chromosomes do not correspond to only one respective distinct parameter in the set of parameters to be optimized. Based on the virtual chromosomes, many new chromosomes are generated. Each new chromosome contains a sequence of values. Each position in the sequences in the new chromosomes corresponds to only one respective distinct parameter in the set of parameters to be optimized. The computer may be configured based on a best new chromosome.

The invention relates to a supervision algorithm management method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring resource use information of a plurality of servers and resource configuration demand information of a plurality of supervision algorithms; determining supervision algorithm configuration information corresponding to each server according to the resource use information of the plurality of servers and the resource configuration demand information of the plurality of supervision algorithms; the supervision algorithm configuration information corresponding to each server is sent to an operation process management and control platform; the supervision algorithm configuration information is used for enabling the operation process management and control platform to call an algorithm in the corresponding server based on the field operation information. The method provides a systematic algorithm management method, and provides a data basis for a follow-up operation process management and control platform to call algorithm execution corresponding to field operation information based on supervision algorithm configuration information corresponding to each server when processing the field operation information.