12 results about "Dynamic Extension" patented technology

This invention relates to an enterprise-level full-stack low-code development platform based on a model-driven and plug-in architecture, belonging to the field of computer technology. It addresses the problems of poor scalability, performance bottlenecks, high security risks, and inflexible deployment inherent in existing platforms. The invention includes: a visual development tool providing a form designer and workflow engine, supporting drag-and-drop development and real-time preview; a plug-in extension module, implemented based on the PF4J framework, supporting dynamic loading, unloading, and isolation of plug-ins; a data integration and management module providing an API gateway, multi-source data integration, and data model management; and an application deployment and maintenance module for one-click automated deployment and log auditing. This invention achieves dynamic functional expansion through a plug-in architecture, optimizes performance in high-concurrency scenarios through a workflow engine, enhances security, and provides more flexible deployment, improving the development and maintenance efficiency of enterprise-level applications. It is suitable for application scenarios such as government approvals.

The invention relates to the field of program development, and provides an iOS system dynamic interaction scene construction method and system. The method comprises the steps that a dynamic class construction tool is developed, page classes are dynamically registered and class members are added when an application runs, and a page class structure capable of being dynamically expanded is obtained; defining a communication protocol between the application and the cloud, obtaining rendering parameters from the cloud based on the communication protocol, assigning the rendering parameters to view attributes of page class instances corresponding to the page class structures, and obtaining a dynamic view module comprising page objects; pre-defining a user operation event, binding the user operation event to a corresponding view of the page object according to event binding configuration of the cloud, and obtaining an event response module; setting a dynamic code execution entry in a system life cycle event of the application so as to execute a JavaScript code issued by a cloud end and obtain a logic execution module; and assembling a rendering system, generating a native UI interface and responding to user interaction operation. The dynamic interface updating of the application is realized, and the user experience is improved.

The application provides a dynamic expansion method of a model and a device thereof, and relates to the technical field of terminal equipment, financial technology and information technology industry. The application determines a target component from a graphical component library, generates a business model based on the target component, structures the business model in a meta model, obtains metadata of the business model, and reflects the metadata to generate a Java object of the business model runtime. The application provides a method for writing a business model through graphing, which can quickly generate a backend Java object while generating a front-end webpage. Compared with a traditional programming code writing method, when an application developer designs an application through a visual method, the threshold is low, the iteration speed is improved, the model function can be dynamically enhanced at runtime, the model method does not need to be defined in advance, the program maintenance cost is reduced, and the expansion capability of the program is improved.

The invention provides a public component development management method. The method comprises the following steps: establishing a code writing specification, a dynamic expansion mechanism, a code annotation specification and a version control strategy for developing a public component; establishing a development team training mechanism and a code review mechanism based on the code writing specification; designing dynamic parameters based on a dynamic expansion mechanism; optimizing the annotation document based on the code annotation specification; and managing the version of each public component in the public component library based on the version control strategy. By establishing a series of management specifications and optimization mechanisms, the development efficiency, quality and maintainability of the public component are improved, and the flexibility and expandability of the public component are enhanced. Through the implementation of the steps, it is ensured that developers can rapidly and easily integrate and call the public components, and meanwhile customized configuration and expansion can be conducted according to specific requirements of the project, so that the development efficiency is improved, the maintenance cost is reduced, and successful implementation of the project is promoted.

The application relates to the technical field of computers and provides a document automatic generation method and system based on a language generation model, which comprises the following steps: in response to generation request information, creating a task scheduling management unit, defining a plurality of initial tasks and creating corresponding task description bodies, and constructing a task association topology based on data dependency relationships; the task association topology is cyclically scanned, schedulable tasks are identified and submitted to a concurrent execution unit for parallel processing, and a language generation model is called to generate content segments; after target task execution is completed, a new task description body is generated based on the output result, corresponding dependency relationships are established and registered to the task association topology, and result-driven incremental topology growth is realized; after all the tasks are completed, the content segments are synthesized and content integration processing is performed, and a target document is generated. The application realizes efficient concurrent generation and dynamic expansion of the structure of the document, and guarantees context consistency.

The class incremental learning method based on multi-granularity distillation and weight alignment comprises the steps that under a given wireless electromagnetic signal data set, an incremental deep learning classification model is initialized, and the model dynamically expands an output layer to adapt to a newly-added class; for a training sample of a current task, inputting the training sample into the model for forward propagation, obtaining a prediction score vector of the model on each category, and calculating cross entropy classification loss for learning discriminant features of a newly introduced category in the task; after training of each task is completed, forward propagation is carried out on all training samples of the current task by using the current model to obtain logs representation of the training samples, and clustering is carried out; in each iteration in the training process, calculating cross entropy loss by using a training sample of a current task, and forming a loss function of a historical task sample; after task training is finished, weight alignment operation is executed, and the difference of new and old category classifiers in weight amplitude is eliminated; and repeating the above steps, performing incremental learning task by task, and continuously performing model parameter updating and memory bank maintenance until all incremental learning tasks are completed.

The present application belongs to the technical field of computer vision, and particularly relates to an open-vocabulary object detection method based on multi-modal fusion and dynamic expansion, aiming to solve the problems of insufficient robustness of visual features, static multi-modal fusion, and poor flexibility of class expansion in the prior art. The method first acquires an image to be detected and text information containing class and scene description, extracts features through a visual encoder and enhances them through a robust noise reduction layer, and then generates a candidate detection frame through a full convolutional network, while extracting text features through a CLIP text encoder. Subsequently, the candidate region features are extracted, dynamically fused through an adaptive multi-modal fusion module to generate cross-modal joint features, and finally the detection frame is screened and the confidence is calculated according to the feature similarity. The model training fuses the InfoNCE loss and the cross-entropy loss to optimize the parameters, and the new class dynamic expansion can also be realized through the analysis of natural language by a generative language model. The present application significantly improves the robustness and accuracy of detection in complex scenes, optimizes the multi-modal fusion efficiency, greatly reduces the cost of new class expansion, and the model meets the real-time detection requirements after acceleration, has good engineering deployment advantages, and can be widely applied in the fields of intelligent monitoring, automatic driving and the like.

The embodiment of the invention relates to the field of data conversion, and provides a multi-source domain-specific language metadata conversion method and computing device.The method comprises the steps that in response to received domain-specific language DSL data, the language type of the DSL data is obtained; based on a first grammar rule corresponding to the language type in a preset grammar rule base, analyzing a semantic structure corresponding to the DSL data; converting the semantic structure into first metadata according to the semantic structure and a preset mapping rule base; and processing the first metadata to generate second metadata conforming to a preset format. Based on the scheme, unified analysis and conversion can be carried out on the multi-source DSL data, the adaptation cost is effectively reduced, and the expansibility and compatibility of the system are improved. And through the modular design of the grammar rule base and the mapping rule base, the analysis capability of dynamically expanding the new type DSL is supported, and the adaptation cost of an external development platform is reduced.

The invention provides a self-evolution model generation and real-time optimization method for an intelligent dialogue system and a related device, and relates to the technical field of artificial intelligence. By establishing a collaborative optimization mechanism, dynamically expanding a model architecture and realizing real-time iterative update, the problems of model update lag, learning mechanism isolation, architecture stiffness and iteration delay are solved, and the method has the advantages of improving the dynamic adaptability, collaborative optimization capability and real-time response efficiency of the system.

According to the low-code platform dynamic expansion system and method based on the standardized plug-in architecture, metadata, interfaces, dependency and authority specifications of the plug-ins are defined through the standardized meta-model, compatibility between the plug-ins and the platform and compatibility between the plug-ins are ensured, and the fragmentation problem of plug-in development is solved. The AI capability engine supports plug-in full-life-cycle intelligent decision making, the development threshold is greatly reduced, and the development and use efficiency is improved. During running, the container adopts a modular framework, dynamic loading / unloading of the plug-in is supported, the platform does not need to be restarted, the platform is guaranteed to be continuously available, and the problem that a traditional plug-in needs to be shut down during loading is solved. Through resource isolation and role-based access control (RBAC) authority definition, the influence on the overall performance of the platform due to the abnormity of a single plug-in is avoided, the unauthorized operation is prevented, and the operation safety of the platform is ensured.

The invention relates to the field of distributed operating systems, and discloses a dynamic expansion-supporting hybrid architecture-based swan-gap system adaptation method, which is technically characterized by comprising a registration verification step, verifying the configuration file of the application module through a preset multi-dimensional verification mechanism to obtain a registration verification result; an operation monitoring step: during the operation period of the application module, acquiring the resource use condition of the application module in real time by using the improved monitoring index set, and generating multi-dimensional monitoring data with a time stamp; a resource demand prediction step: obtaining the multi-dimensional monitoring data, generating a prediction result based on the time sequence prediction model, the prediction result reflecting the resource demand of the application module in the future time period; and a dynamic resource adjustment step: according to the prediction result and the registration verification result, dynamically adjusting the resource allocation of the application module by using a preset dynamic resource adjustment strategy.