13 results about "Process deployment" patented technology

“Prevention is better than cure” aptly applies to the field of cybersecurity. However, state-of-the-art integration systems generally perform late-stage scanning that is reactive, rather than proactive. This can be particularly dangerous in a low-code environment, in which the integration processes may be constructed by novice users without expertise in mitigating cybersecurity threats. Accordingly, embodiments utilize artificial intelligence to proactively detect and resolve cybersecurity threats during the design stage of integration processes. For example, a generative model may be used with crowd-sourced integration configurations to generate synthetic integration data that are targeted towards specific cybersecurity threat endpoints. This synthetic integration data may be input to an integration process, during integration testing within a test environment, to detect and resolve cybersecurity threats, prior to deployment of the integration process to a production environment.

The application discloses an application flow deployment method and device, computer equipment, a computer readable storage medium and a computer program product, and particularly relates to the field of big data. The method comprises the following steps: obtaining configuration information, the configuration information carrying a template label; obtaining an initial flow template corresponding to the template label, and generating an initial application flow according to the initial flow template and the configuration information; the initial application flow comprises a plurality of task nodes; each task node in the initial application flow is configured according to the configuration information to obtain a target application flow; the target application flow is subjected to integrity verification and business rule verification, and the target application flow is deployed when the verification is passed. The process of generating the initial application flow comprises the following steps: obtaining the initial flow template corresponding to the template label from a template library; and updating the initial flow template according to a flow structure in the configuration information to obtain the initial application flow. The method can efficiently realize automatic deployment.

The application provides a method and device for creating an intelligent application, electronic equipment and a medium, which can be widely applied in the field of computer technology. The method for creating an intelligent application comprises: obtaining a target requirement of a target object, and collecting training data from a target data source based on the target requirement; processing the training data through a model training scaffold of a machine learning platform to obtain preprocessed data; establishing a pre-training model through the model training scaffold of the machine learning platform, and training the pre-training model according to the preprocessed data to obtain a target pre-training model; storing the target pre-training model in a product library; and creating an intelligent application based on the target pre-training model in the product library through an agent application scaffold or an inference service scaffold of a technical platform. The application integrates a machine learning platform and a technical platform to realize the whole-process deployment of an intelligent application, which is conducive to improving the efficiency of intelligent application development.

This specification discloses a method, apparatus, storage medium, and electronic device for processing deployment tests. The method includes: monitoring service strategy experiment status data and constructing and managing a wide table of user traffic deployment experiment records; upon receiving a target strategy deployment request, obtaining a target set of test user entities, performing pre-contact conflict detection based on the wide table, and removing and isolating entities overlapping with already run test experiments; during the test run phase, periodically monitoring the wide table and the target set, performing runtime conflict detection, and triggering anomaly warnings; after the experiment, performing data overlap detection on the experimental group and the control group, and generating an experimental data independence test report based on the overlap information. This specification embodiment achieves automated conflict management throughout the entire test task lifecycle, effectively avoiding traffic pollution between concurrent experiments on the platform, and improving the independence of experimental data and the confidence of the final results.

The present application relates to the technical field of hardware acceleration and artificial intelligence inference deployment, and discloses a method and system for deploying a convolutional neural network after training, comprising S1 environment initialization; S2 model training and freezing; S3 frozen graph precision evaluation; S4 quantization and quantization precision evaluation; S5 compilation to generate a deployment model; S6 deployment package generation and DPU activation; S7 inference verification and performance statistics. The method for deploying the convolutional neural network after training can decompose the CNN deployment process after training into a series of ordered steps such as environment initialization, model training, computation graph freezing, frozen graph precision evaluation, quantization calibration, quantization graph precision evaluation, DPU model compilation for target FPGA platform, deployment package generation, DPU activation and inference verification, by constructing a standardized and scriptable full-process deployment link, and each link is executed in series through a scriptable way, and the link effectively reduces the probability of human operation errors.

The invention discloses a trusted execution environment remote authentication method and system suitable for a federated learning scene, and relates to the technical field of federated learning and trusted execution environments. The method comprises the following steps: carrying out equipment registration according to a registration request of trusted execution environment equipment; performing remote authentication operation on the federated learning administrator based on the registered trusted execution environment, specifically comprising the following steps: performing identity authentication and trusted execution environment authentication on the federated learning administrator, performing a remote certification process and a key negotiation process at the same time, and deploying a federated training model and an aggregation model; and remote authentication is carried out on other participants participating in the federated learning task based on the federated training model and the aggregation model, and the method specifically comprises the following steps that trusted execution environment registration and information uploading operation are carried out on the participants, and the participants execute the federated learning task according to the deployed federated training model and the aggregation model. According to the invention, an efficient remote authentication process can be realized in a federated learning multi-participant scene.

An agent management platform for providing a runtime environment for the execution of agents can be used to manage process deployments. The agents can be configured to perform specific tasks on software applications within defined objectives. Using an obtained execution request, the agent management platform can associate agents with one or more nodes by comparing node constraints against agent constraints. The nodes can include computers equipped with containers that provide isolated runtime environments for agent execution. The agent management platform can instantiate agents in corresponding containers on the computers of associated nodes, and / or execute the instantiated agents to perform corresponding tasks.

This invention provides a method and system for AI model copyright protection and trusted reasoning, comprising the following steps: The model provider encrypts the AI ​​model using a first key and sends the first key to a verification service; the verification service configures access control conditions for the first key, the access control conditions being associated with the hardware identity identifier and platform trusted state requirements of the target deployment platform; when the deployment platform requests to load the encrypted AI model, a remote proof and key release process is executed: The deployment platform generates proof information containing its hardware identity identifier and current platform state measurement value; the proof information is sent to the verification service; the verification service verifies the authenticity of the proof information and determines whether the hardware identity identifier and platform state measurement value comply with the access control conditions; if they comply, the verification service encrypts the first key using an encryption key corresponding to the deployment platform's hardware security module and then issues it; the deployment platform decrypts to obtain the first key.

The application discloses a traffic research and judgment analysis platform management method based on an Activiti process engine, and comprises the following steps: S1. process modeling: using a process modeling tool provided by the Activiti process engine, defining a workflow model related to people, vehicles, roads and business processes; S2. process deployment: deploying the defined workflow model into the Activiti process engine; S3. process execution and monitoring: triggering the corresponding workflow according to actual needs and monitoring in real time; S4. process optimization and iteration: according to feedback and data analysis results in actual application, optimizing and iterating the workflow model. The application applies the Activiti process engine to the traffic research and judgment analysis platform, can significantly improve the management efficiency, accuracy and automation level of the platform, optimizes the business process, enhances the scalability and flexibility of the platform, and greatly improves the security and reliability of data.

Conventional troubleshooting for integration processes in an integration platform is inefficient and requires significant expertise. Accordingly, an error resolution model is disclosed. The error resolution model may be operated to predict an error resolution, based on the current design (e.g., lineage) of an integration process, during construction of that integration process (e.g., on a virtual canvas). A generative language model may also be used to produce dialogs for the error resolutions. This enables the efficient troubleshooting and resolution of errors in an integration process, prior to that integration process being deployed and executed, and without requiring significant expertise.

Conventional troubleshooting for integration processes in an integration platform is inefficient and requires significant expertise. Accordingly, an error resolution model is disclosed. The error resolution model may be operated to predict an error resolution, based on the current design (e.g., lineage) of an integration process, during construction of that integration process (e.g., on a virtual canvas). A generative language model may also be used to produce dialogs for the error resolutions. This enables the efficient troubleshooting and resolution of errors in an integration process, prior to that integration process being deployed and executed, and without requiring significant expertise.

This invention discloses a low-code artificial intelligence software engineering system for biometric recognition, belonging to the field of artificial intelligence technology. It includes a standardized biometric access module, a low-code visual orchestration module, a lightweight AI model adaptation module, an automated process deployment module, a performance monitoring and iterative optimization module, and a permission security control module. By standardizing the processing of multiple types of biometric data and combining it with low-code visual orchestration to construct the recognition process, it performs lightweight customization and hyperparameter tuning of pre-trained models, achieving the fusion compilation of engineering scripts and models and cross-platform automated deployment. Simultaneously, it monitors system performance in real time and completes self-supervised iterative optimization of the model, supplemented by end-to-end permission security control. This invention lowers the development threshold of biometric recognition AI systems, improves the system's cross-scenario adaptability and deployment efficiency, realizes full lifecycle engineering management, ensures biometric data security, and can be quickly adapted to multiple application scenarios such as access control and attendance.