264 results about "Deployment environment" patented technology

Systems and methods for machine learning model retraining is described. In one example, a system includes a computing device that is configured to determine an anomaly score from a first input of a featurized data set for a machine learning model of a deployment environment. The computing device is configured to determine a feature correlation score for the machine learning model based at least in part on a second input of a featurized historical data set for the machine learning model. A model retraining frequency time period for the machine learning model is determined based at least in part on the feature correlation score and the anomaly score.

A reliable reasoning scheduling method based on an edge hybrid expert large model relates to the field of computers, and comprises the following steps: deploying a MoE large model to an edge network in a distributed edge manner; malicious node behavior modeling is carried out, distributed cooperative reasoning is simulated, and an auto-encoder input vector space set V and a reasoning task state set S are generated respectively; performing abnormal node detection neural network training to obtain calculation reliability of each edge node; performing deep reinforcement learning neural network training to obtain an expert selection strategy; and performing reliable reasoning scheduling of the edge MoE large model based on the trained deep reinforcement learning neural network. According to the method, the service quality and privacy protection advantages of distributed edge deployment are utilized, according to an original routing result of a large model gating function to an expert network, the node reliability and real-time load in an actual dynamic deployment environment are combined, the reasoning process is effectively optimized, and the reasoning efficiency is improved. Therefore, the cooperative reasoning efficiency and reliability of the MoE large model under distributed edge deployment are improved.

The invention discloses an intelligent file service system generation method based on a low-code development mode. The method comprises the following steps: constructing a file service core data model, defining file metadata fields, a storage structure and an association relationship, and forming a standardized data architecture. According to the method, the core data model, the visual page component library and the drag-and-drop workflow template are constructed, linkage binding of data-component-process is achieved in combination with a low-code engine, and the efficiency bottleneck of a traditional development mode is broken through; through a three-way real-time synchronization mechanism, real-time hot repair and dynamic branch judgment, page layout and flow rule adjustment can take effect immediately through visual configuration, and the problem of dynamic demand response lag is solved; through the intelligent resource scheduling logic compiled through environment adaptation, a differentiation scheme is generated for different deployment environments, the configuration difference influence under a private cloud and server mixed deployment scene is remarkably reduced, the deployment failure rate is greatly reduced, and the system online stability is improved.

The invention relates to the field of waterproof sockets, and discloses a multi-scene adaptive intelligent waterproof socket remote regulation and control system and method, and the method comprises the steps: obtaining multi-source sensing data of a socket deployment environment, and constructing a multi-scene adaptive input data set in combination with a parameter fusion strategy and a signal interference compensation mechanism; environment feature recognition is carried out on the multi-scene adaptive input data set, and typical application modes in different scenes are extracted by using a socket state tag extraction algorithm; based on the state mapping relation model, a remote regulation and control strategy optimization algorithm is adopted to dynamically analyze the operation state of the socket, and the matching degree of the current operation mode and the environment scene is identified; an intelligent judgment mechanism is introduced, pre-execution verification is carried out on the multi-scene regulation and control suggestion set, and potential control conflicts are recognized in combination with historical operation data and real-time control instructions; and performing instruction issuing management on the remote control behavior of the socket based on the optimized execution strategy. The invention has the advantage of improving the universality of the intelligent waterproof socket.

The invention relates to the technical field of data transmission of the Internet of Things, in particular to a self-adaptive data compression and transmission method of a low-power-consumption wide-area Internet of Things. According to the method, a three-layer collaborative architecture comprising an equipment layer, a fog computing layer and a cloud computing layer is constructed, data preprocessing and feature analysis are performed on the equipment layer, and data types, entropy values and repeatability information are extracted; the fog calculation layer selects an optimal compression algorithm based on a multi-dimensional decision engine, and reduces redundancy through spatial-temporal correlation analysis and data aggregation; the cloud computing layer collects compression performance data, adopts reinforcement learning and federal learning to train a global optimization model, and dynamically issues strategy parameters; and the fog node adaptively adjusts a compression strategy in combination with the system state to realize the optimal balance between the compression ratio and the reconstruction precision. The method is suitable for field deployment environments with limited electric quantity and network, has the advantages of low power consumption, high efficiency, strong adaptability and the like, and can be widely applied to Internet of Things scenes such as remote monitoring, smart energy and the like.

The invention relates to the technical field of computers, in particular to a dynamic configuration initialization and version synchronization method and system for multi-agent service, and the method comprises the following steps: obtaining a database update lock, constructing a database migration script dependency relationship, determining a current version of a database mode, upgrading and deploying the database mode, and verifying agent versions. The method has the beneficial effects that a rigorous database updating lock control mechanism, perfect migration script dependency chain verification, agent version management, port availability detection, idempotence and transaction safeguard measures are introduced, so that the processes of database structure upgrading and agent version initialization in a distributed deployment environment can be effectively supported; and the data consistency, the operation stability and the service continuity of the system are ensured.

In one embodiment, a device identifies a plurality of tasks that a base machine learning model is able to perform. The device receives, via a user interface, a request to generate a specialized model to perform a particular task for deployment to a target deployment environment. The device uses knowledge distillation on the base machine learning model to train the specialized model to perform the particular task based on at least one of the plurality of tasks. The device causes the specialized model to be deployed to the target deployment environment.

The invention discloses a model deployment method and device combined with hardware deployment, and the method comprises the steps: constructing a hardware topological graph corresponding to equipment nodes according to the equipment nodes in a target model deployment environment and a connection relation corresponding to each equipment node; analyzing a dependency relationship of a target operator in the target model to generate a model calculation graph corresponding to the target operator; calculating a target matching value between the equipment node and the target operator by using a preset evaluation algorithm according to the hardware topological graph and the model calculation graph; and determining a deployment scheme corresponding to the target model according to the target matching value and a preset grading strategy. By sensing a dynamic matching mechanism of a hardware state and a model calculation graph in real time, an optimal deployment strategy is automatically generated, and adaptive optimization is triggered when hardware topology or resource conditions change, so that real-time collaborative adaptation of a model segmentation strategy and a heterogeneous environment is realized, and the deployment automation level and the system elasticity are remarkably improved.

The invention provides a cloud computing platform automatic deployment method, electronic equipment and a storage medium, and belongs to the technical field of cloud platforms. The method comprises the following steps: acquiring deployment host information corresponding to a deployment host, target host information corresponding to a target host and component configuration information; filling a preset deployment environment detection instruction template according to the deployment host information and the target host information, and generating a deployment environment detection script; performing environment detection on the deployment host and the target host based on the deployment environment detection script to obtain a deployment environment detection result; in response to the environment detection result, indicating that the deployment host and the target host meet a preset deployment condition, filling a preset component deployment instruction template according to the component configuration information, and generating an automatic deployment script; and executing the automatic deployment script to enable the deployment host to deploy the target cloud computing platform at the target host. According to the invention, automatic deployment of the cloud computing platform can be realized, and the deployment efficiency of the cloud computing platform is improved.

The invention discloses a visual arrangement system and method for urban rail cloud simulation test resources, and aims to assist a user to conveniently and efficiently arrange a simulation test service scene, convert the service scene into a template, namely a reusable service scene architecture script, and enable a scheduling engine to schedule the simulation test service scene according to template data information. The method comprises the following steps: respectively docking with an on-cloud platform, an under-cloud system and a simulation interface, establishing a simulation test scene instance of a service system from the on-cloud to the under-cloud, and enabling a user to carry out the simulation test work of the service system on the scene instance. According to the method, a certain resource can be selected as a template to be stored, when needed in the follow-up simulation test process, a template file is flexibly called, the environment is pulled up and deployed through a scheduling engine, the problem that the resources of the test environment are limited is solved, and the convenience and efficiency of cloud simulation environment scene construction are improved.

The invention discloses a heterogeneous document structured data extraction system and method based on multi-modal fusion, and the method comprises the following steps: S1, receiving a heterogeneous document, and carrying out the preprocessing; s2, extracting visual and semantic modal information based on the visual backbone network and an OCR module, and fusing spatial coordinates, text content and layout features; s3, receiving a dynamic target mode Schema, executing field-level semantic matching, and calculating semantic similarity; s4, performing consistency verification, verifying and filling the field values, and correcting the associated fields; s5, converting the format into a data file in a specified format, and reserving fields to be mapped with an original document semantic block; s6, automatically switching a cloud mode and a local mode according to a deployment environment; and S7, outputting the data file of the target field. According to the method, high-precision and low-delay structured data extraction of the heterogeneous format document can be realized, and the automatic processing efficiency and the data credibility are improved.

The invention relates to a dynamic environment-oriented laser SLAM (Simultaneous Localization and Mapping) method and system and a medium. Comprising the following steps: training and deploying a target detection four-dimensional neural network; performing environment initialization and data processing, and constructing an initial dense point cloud map; forward and reverse propagation is carried out by using IMU data, and the current state of the robot is accurately estimated; performing point cloud registration to update the map; identifying and marking feature points of the dynamic target; updating the dynamic or static probability of voxels in the map according to the dynamic target detection result; and outputting the optimized map and the pose estimation of the robot. According to the method, the understanding of the dynamic environment is updated by adopting the voxel belief model, the information of the dynamic object is effectively fused into the map, meanwhile, the accurate representation of the static environment is reserved, and the accuracy and robustness of positioning, map construction and dynamic object recognition of the SLAM system in the dynamic environment are improved. And through an optimized data processing flow, efficient calculation performance and real-time performance are ensured.

The invention provides a software deployment method which can be applied to the technical field of artificial intelligence. The software deployment method comprises the following steps: in response to received code data uploaded by a user, analyzing the code data to obtain a dependency relationship among software items in the code data, and determining a priority corresponding to the dependency relationship; according to the priorities, the software projects are constructed in sequence to generate target software; according to the deployment requirement of the target software, searching from a pre-constructed knowledge base to obtain configuration information matched with the deployment requirement, and configuring a target deployment environment according to the configuration information; and extracting deployment parameters from the target deployment environment, and deploying the target software to the target deployment environment according to the deployment parameters. The invention further provides a software deployment device and equipment, a medium and a program product.

The invention relates to the technical field of software engineering management, and discloses a computer software development information management system and method. According to the method, real-time multi-dimensional association among four entities, namely a demand entity, a code entity, a test entity and a deployment entity is realized by constructing a dynamic knowledge graph center, and an end-to-end closed-loop control system is formed; a full-link traceable network is dynamically established through explicit binding, implicit capturing and cross-layer traceability mechanisms, demand omission marks are automatically repaired, and accurate attribution of environmental problems is achieved; manual operation blind areas are eliminated through automatic means such as developing and submitting forced association, automatically freezing branches when test fails and deploying environment consistency verification; based on a risk thermodynamic diagram of full-link topology penetration and fusion defect density and change frequency, active early warning of a high-risk module is realized, and controllability and quality assurance capability of a software development process are remarkably improved.

The invention relates to the technical field of data center machine room energy conservation, and discloses a machine room cold channel closed management method and device and terminal device.The method comprises the following steps that the state of a cold channel is obtained by deploying environment monitoring nodes, a heat-flow coupling model and a stability function are constructed, a feedback controller is designed to generate an air door and refrigeration control instruction, and a heat-flow coupling model is established; linkage adjustment of the air velocity and the temperature is achieved, closed-loop control is constructed, and efficient and energy-saving management of a cold channel is achieved; the device comprises an environment monitoring module, a state modeling module, a stability evaluation module, a state feedback control module and an execution linkage module. Through state fusion updating, a component control structure, a stability function driving strategy and a linkage execution mechanism, fine adjustment and efficient cooperative control of the temperature and airflow of the cold channel are achieved, the adjustment stability and response efficiency are improved, energy consumption is remarkably reduced, and the problems that an existing system is poor in control precision, poor in linkage performance and the like are solved.

The invention provides an accelerator card hybrid deployment method and device and electronic equipment, and the method comprises the steps: carrying out the segmentation of an overall calculation graph, and obtaining a segmented calculation graph; inputting the calculation characteristic index of the segmentation calculation graph into a performance prediction model to obtain performance prediction data, output by the performance prediction model, of the segmentation calculation graph executed by various acceleration cards; and based on the performance prediction data of the segmentation calculation graph, carrying out acceleration card hybrid deployment on the target task. According to the method provided by the invention, the integral calculation graph is segmented to obtain the segmented calculation graph; inputting the calculation characteristic index of the segmentation calculation graph into a performance prediction model to obtain performance prediction data, output by the performance prediction model, of the segmentation calculation graph executed by various acceleration cards; based on the performance prediction data of the segmentation calculation graph, accelerator card hybrid deployment is performed on the target task, so that accurate accelerator card hybrid deployment is realized, and then the execution efficiency of executing the target task is improved and the power consumption required by execution is reduced in a hybrid deployment environment.

The embodiment of the invention discloses a security deployment method and device for enterprise-level applications and a medium, and relates to the technical field of security deployment, and the method comprises the steps: obtaining application topological graph data generated by a preset visual development platform, so as to split the application topological graph data into an atomized micro-service unit set through a quantization deconstruction engine, generating a micro-service dependency tree; scanning the target deployment environment, extracting an environment feature fingerprint of the target deployment environment, and generating an environment adaptation layer code according to the environment feature fingerprint; injecting an environment adaptation layer code into each atomized micro-service unit to generate an environment perception type micro-service package, and constructing a cross-cloud communication link matrix based on the micro-service dependency tree; and based on the environment-aware micro-service packet and the cross-cloud communication link matrix, executing a deployment verification test in the sandbox environment, generating a deployment verification report, and according to the deployment verification report, distributing the environment-aware micro-service packet to a target deployment environment node.

In various examples, a declarative configuration format defines a machine learning pipeline using nodes and edges that define the data flow between those nodes. Partial pipelines may be defined independently and referenced by multiple full pipelines. The configuration format may allow for constructing applications for a range of hardware and deployment environments. The system may support both graphical and programmatic interfaces for defining, modifying, and visualizing pipelines. The machine learning pipelines may be implemented on or interact with an underlying machine learning pipeline framework that provides runtime objects and a plugin architecture. The configuration data may be used to generate an application that includes the processing components and corresponding runtime dependencies of a machine learning pipeline in a containerized image.

A method, apparatus, and non-transitory computer-readable medium for packaging and deploying applications using virtualized execution units (VEUs). A development apparatus instantiates a development environment, captures changes to identify build elements, and packages the build elements into a VEU. An execution apparatus receives and parses the VEU to obtain build elements, determines corresponding deployment actions, and applies these actions to instantiate the application in a deployment environment, automating application deployment while preserving state.

The invention discloses a platform authentication and configuration loading method and device, electronic equipment and a medium, and the method comprises the steps: generating a platform unique identifier based on a platform name of a target platform, and obtaining configuration information of the target platform in response to a configuration demand of a target object; performing multiple encryption on the platform unique identifier and the configuration information to generate a license file; the license file comprises a first public key, a key ciphertext, an identification ciphertext, a verification ciphertext and an information ciphertext; transmitting the license file to a target deployment environment, so that the target deployment environment performs multi-segment decryption on the key ciphertext, the identification ciphertext and the information ciphertext based on the first public key and performs consistency verification based on the verification ciphertext to obtain decryption configuration information; and carrying out platform initialization loading according to the decryption configuration information. The embodiment of the invention can safely and efficiently realize platform authentication and configuration loading, and can be widely applied to the technical field of data processing.

The invention discloses a multi-model intelligent management scheduling method and system. The method comprises the following steps: receiving a to-be-registered model and corresponding metadata, and executing integrity check to generate an integrity check result; generating a model identifier for the to-be-registered model according to an integrity check result, and storing the model identifier into a model storage warehouse; receiving a deployment request, and extracting a target model file and target metadata according to the deployment request; selecting a target computing node according to the target metadata; configuring a running environment according to a deployment environment requirement, and combining the target model file with the running environment to obtain a container mirror image; creating a containerized operation instance based on the container mirror image, and configuring a network access endpoint; predicting the page view of the network access endpoint to obtain predicted page view data; and adjusting the number of the containerized operation instances according to the predicted page view data and the system resource use data. By implementing the technical scheme, the problems of hysteresis and lack of perspectiveness of resource adjustment based on a real-time system load are solved.

Container image optimization is provided for a deployment environment. An indication of a plurality of container images to be deployed in a deployment to provide a plurality of services within a computing environment is obtained. The plurality of container images of the deployment has a total layer size of a plurality of image layers of the plurality of container images. At least one container image of the plurality of container images is reconfigured to adjust the total layer size for the deployment. The reconfiguring is based on one or more deployment characteristics of the deployment and includes replacing at least one image layer of a container image of the at least one container image with one or more other image layers configured based on the plurality of container images to be deployed.

Methods and apparatus to implement post-provisioning tasks are disclosed. An example apparatus comprising memory, instructions, and programmable circuitry to be programmed by the instructions to obtain metadata associated with a post-provisioning task, the post-provisioning task to modify a plugin, the plugin to provide a capability to a cloud resource of a computing platform provider, the metadata represented in accordance with a first file format, transform the metadata from the first file format to a second file format, the second file format compatible with the plugin, and register the post-provisioning task in a deployment environment provided by the computing platform provider.

The invention provides a multi-language compatible algorithm scheduling platform and method, and the platform comprises a data collection module which is used for receiving basic information and algorithm files of different language algorithms based on an API interface, and converting the algorithm files into mirror image files according to the basic information based on a preset method; the algorithm deployment module is used for carrying out environment configuration on the algorithm container based on the basic information, carrying out isolated deployment on the mirror image file in the algorithm container based on an environment configuration result, and carrying out multi-dimensional management on different language algorithms subjected to isolated deployment; and the algorithm scheduling module is used for receiving a service demand based on the multi-dimensional management result, locking a target language algorithm based on the service demand, and performing scheduling operation on the target language algorithm based on an API interface. According to the method, different language algorithms can stably and reliably run in a deployment environment, the reliability of cross-language deployment is improved, and the reliability, stability and accuracy of multi-language algorithm scheduling are improved.

Methods and systems for managing services (e.g., the development, deployment, and configuration of services) are disclosed. User intent associated with policy related configurations for computer implemented services are directly and automatically translated into a set of user intended network policy and rules to be applied (e.g., enforced) by the computer implemented services without any need for the user to understand how the policies will be configured and implemented by the service deployment environment.

The invention discloses a DevOps-based dynamic component security management method and system, and belongs to the technical field of information security detection. Component information data and vulnerability information data are matched by using a matching engine, potential security vulnerabilities are identified, and the security vulnerabilities of the component information data and the vulnerability information data are matched; according to the severity of the vulnerability, the deployment environment of the project, the involved user data volume and the involved sensitive data volume, risk assessment is carried out on the security vulnerability, automatic rating and classification are carried out on risk levels, the priority of the security vulnerability and a corresponding repair strategy are confirmed, a risk report is generated, and the risk report is displayed or exported through a user interface. According to the dynamic component security management method and system based on DevOps disclosed by the invention, potential security vulnerabilities can be accurately identified, automatic risk rating is carried out, vulnerability repair priorities can be ranked according to project risk levels, and later project vulnerability management is facilitated.

The invention provides a network security event identification model deployment method and system in the technical field of network security and machine learning crossing. The method comprises the following steps: S1, acquiring historical monitoring data to construct a source domain data set; s2, training the created security event identification model through the source domain data set, and dynamically pruning and quantifying the security event identification model in the training process; s3, performing knowledge distillation on the security event identification model; s4, acquiring real-time monitoring data of a to-be-deployed environment to construct a target domain data set, and performing data drift compensation training on the security event recognition model after knowledge distillation based on the target domain data set and the source domain data set; and S5, carrying out migration deployment on the security event identification model after data drift compensation training through a containerization technology and a shadow deployment technology. The method has the advantages that the flexibility and adaptability of network security event identification model deployment are greatly improved, and the deployment cost is reduced.

The invention discloses a lightweight model transfer learning method based on knowledge distillation, and belongs to the technical field of artificial intelligence security, and the learning method comprises the steps: S1, constructing a security initialization module, and deploying a teacher model in a preset cloud server; s2, a lightweight student model is generated, and a preset processing module receives the teacher model and performs model compression processing on the teacher model; s3, knowledge migration safety protection is carried out, and the lightweight student model is trained; s4, deploying environment security binding, and carrying out binding verification on the trained lightweight student model and preset target operation equipment; and S5, full-cycle audit tracking is carried out, a traceable audit management module is established, and a full-process operation log from teacher model compression to lightweight student model deployment operation is recorded. Through the steps S1-S5, structural information leakage of the original teacher model is avoided, and dynamic permission control is performed on the teacher model in a deployment link.

The invention discloses a Web-based visual editing method, belongs to the technical field of industrial configuration, and aims to solve the problems of complex deployment of traditional configuration software and insufficient Web scheme performance. The method comprises the following steps: constructing a front-end architecture (Vue. Js + Konva.js) containing a graphic element module, a canvas management module and the like; the primitive adopts JavaScript prototype chain inheritance + SVG / Canvas double-layer rendering, and a canvas is divided into four layers and is provided with a quadtree index; an editor and an expression engine are generated by using a factory pattern for attribute configuration; page management supports state capture recovery; real-time communication is based on WebSocket + PubSub, and data aggregation and incremental updating optimization are combined. Cross-platform zero deployment and Web environment near-desktop performance are achieved, editing efficiency is improved, industrial protocols such as OPC UA are compatible, and the method is suitable for the field of energy and intelligent manufacturing.

An MLOps extension method and apparatus in a multi-cloud native environment, and a medium, relating to the technical field of multi-cloud heterogeneity. The method comprises: receiving a user-defined infrastructure template, configuration parameters, and access credentials of different cloud platforms, and collecting cloud computing resources from a plurality of clouds; putting model code and a data processing script into a version control system, and uploading data to a data storage service; pre-allocating computing resources and storage resources, and installing and verifying all dependencies in a pre-deployment environment; creating isolated sandbox environments on the cloud platforms to run comprehensive simulation tests; packaging models and applications, and deploying same in production environments on the cloud platforms; configuring a monitoring tool to monitor the performance of the models and the applications in real time, and on the basis of monitoring data, optimizing model parameters and application configurations. The present invention can reduce incompatibility and failures in a multi-cloud environment, and improve the reliability and efficiency of model services and continuous application deployment.