Visual Model Invocation Workflow for Stable Application Creation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developers face challenges in simplifying and accelerating the application development process, particularly in integrating machine learning models, requiring deep understanding of model operations and configuration, which hinders widespread participation and efficiency.
Innovation Solution
A method and apparatus for application creation that involves selecting a model invocation function block, configuring it through an interface, and creating an application based on a workflow, reducing the complexity of model invocation and enhancing stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If developers directly integrate machine learning models into applications, then application functionality and intelligence are improved, but development complexity and difficulty increase
Solution Approach 1:
The patent introduces a model invocation function block as an intermediary component between the application workflow and machine learning models. This function block abstracts the complex model integration process into a standardized interface, allowing developers to invoke models without directly dealing with their internal complexities, thereby reducing development complexity while maintaining application functionality
Solution Approach 2:
The patent segments the application development process into distinct functional blocks, including model invocation function blocks. By dividing the complex integration task into modular, manageable segments with clear interfaces, developers can work with simplified components rather than monolithic complex systems, reducing overall development complexity
2Manufacturing precision
If developers deeply understand model operations and configuration, then model integration accuracy and performance are improved, but development time and skill requirements increase
Solution Approach 1:
The model invocation function block is designed to automatically handle model configuration and invocation processes. The system self-manages the complex interactions between application workflows and machine learning models, reducing the need for developers to deeply understand model operations while maintaining integration accuracy, thereby reducing development time and skill requirements
3Ease of operation
If traditional application development methods are used, then development simplicity is maintained, but application intelligence and model integration capability are limited
Solution Approach 1:
The model invocation function block serves multiple functions: it acts as a workflow component, a model interface, and a configuration manager. This universal component can be used across different applications and model types, maintaining development simplicity through standardized usage while enhancing model integration capability through its multi-functional design
Data Source
AI summary
The embodiments of the disclosure provides a method, apparatus, device and storage medium for application creation. The method includes: selecting, for a model invoking node in a plurality of processing nodes of a target application to be created, a model invocation function block for invoking a first model by the model invoking node from a plurality of function blocks; presenting a configuration interface for the model invocation function block, in response to a configuration request for the model invocation function block; receiving configuration information for the model invocation function block via the configuration interface; and creating the target application based on the configuration information and a workflow composed of the plurality of processing nodes. By the configurability and visualization of the model invocation, the development difficulty of the application with the model invocation capability may be reduced, and the stability of model outputs may be improved.