Learning Model Generation via Unified Interface and Pre-Existing Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating a trained learning model is challenging due to the need for specialized knowledge and time-consuming programming, especially when adapting existing learning models to new training data sets.
Innovation Solution
A learning model generation device and method that displays multiple learning models with a unified interface, allowing users to select and subject them to machine learning, eliminating the need for programming by using pre-existing models and generating training data through graphical user interfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If an existing learning model is adapted to a new training data set, then the model can be reused and development time is reduced, but programming knowledge and time are still required to code the adaptation process
Solution Approach 1:
The system performs automatic model adaptation through self-service mechanisms. The model selection unit automatically selects appropriate learning models from the library, and the training data generation unit automatically generates and adapts training data without requiring user programming. The system serves itself by autonomously completing the adaptation process based on predefined algorithms and user inputs through GUI only.
Solution Approach 2:
The system introduces an intermediary layer between the user and the machine learning process. This intermediary includes the model selection unit, training data generation unit, and machine learning execution unit, which together mediate the adaptation process. The user interacts only through the GUI, while the intermediary handles all programming-intensive tasks of model selection, data generation, and training execution.
2Manufacturing precision
If a learning model is constructed from scratch based on machine learning algorithms, then the model can be precisely tailored to specific needs, but significant programming knowledge and time are required
Solution Approach 1:
The system performs preliminary actions by pre-storing multiple learning models in the model library with different architectures and characteristics. This preliminary preparation allows users to select from pre-configured models rather than constructing from scratch, maintaining customization capability while eliminating the need for programming. The models are pre-adapted to various data types and tasks in advance.
Solution Approach 2:
The system enables model customization through parameter changes rather than structural reconstruction. Users can select from pre-built models and adjust parameters through the GUI, allowing precise tailoring to specific needs without programming. The training data generation unit also applies parameter changes to adapt data formats and characteristics to match selected models.
3Adaptability or versatility
If programming is required for model generation and adaptation, then flexibility and control are improved, but the process becomes inaccessible to non-programmers and time-consuming
Solution Approach 1:
The system replaces the mechanical programming system with a graphical user interface-based interaction system. Instead of requiring users to write and execute code, the system provides visual interfaces for model selection, parameter configuration, and data input. This substitution maintains full flexibility and control while making the system accessible to non-programmers.
Solution Approach 2:
The system achieves universality by designing a multi-functional platform that handles model selection, data generation, training execution, and result evaluation through a unified GUI. This universal interface serves multiple purposes (model construction, adaptation, training, evaluation) without requiring different programming skills, making the system accessible to users with diverse backgrounds while maintaining versatility.
Data Source
AI summary
A learning model generation device includes a model information display that displays a plurality of learning models being released on the Internet and having a same interface in a selectable manner, a model selector that displays at least one learning model among the plurality of displayed learning models, and a trainer that subjects the selected learning model to machine learning such that the selected learning model learns training data, and generates a trained model that has been trained.