Source Model Combination for Low-Data Target Model Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning models struggle to achieve high accuracy with limited data, and existing transfer learning methods are inefficient in model development and require significant preparation and calculation costs.
Innovation Solution
A model generation system that combines trained models, equations, and inequalities from a source model database to create a target model, reducing the need for additional learning data and calculation costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If transfer learning is used to reduce data preparation and calculation cost, then cost efficiency is improved, but model development efficiency remains insufficient due to difficulty in effectively combining multiple trained models
Solution Approach 1:
The patent combines multiple trained models (source models) into a single target model by integrating their inputs and outputs. The model generation unit merges first source model with first output and second source model with second output, creating a unified model structure that leverages multiple existing models while reducing redundant calculations and data preparation costs.
Solution Approach 2:
The system creates a universal model generation approach that can handle multiple source models with different inputs and outputs. The model generation unit is designed to accommodate various combinations of source models, making the system adaptable to different learning scenarios and data types while maintaining consistent integration methodology.
2Measurement precision
If multiple trained models are combined to improve prediction accuracy, then prediction accuracy is improved, but the complexity of model integration and data association increases
Solution Approach 1:
The patent segments the model integration process into distinct functional units: a database search unit that identifies suitable source models, a combination determination unit that validates input-output associations, and a model generation unit that performs the actual integration. This segmentation reduces overall complexity by breaking down the integration task into manageable steps.
Solution Approach 2:
The system introduces an intermediary association verification mechanism that checks whether inputs of source models can be properly associated with outputs of other source models. This intermediary validation step ensures compatibility before integration, preventing complex and erroneous model combinations while maintaining high prediction accuracy.
3Loss of time
If existing trained models are reused to reduce learning time, then development time is reduced, but the ability to generate new models with specific requirements is limited
Solution Approach 1:
The system dynamically selects and combines source models based on the specific requirements of the target model. The database search unit identifies appropriate source models from the stored collection, and the combination determination unit adapts the integration strategy according to the required inputs and outputs, enabling flexible customization while reducing learning time through reuse of existing models.
Data Source
AI summary
A model generation system includes: a source model database 12 that stores a source model; and a model generation unit 11 configured to generate the target model using the source model searched from the source model database. The model generation unit includes a database search unit configured to search for a first source model 31 including an output of the target model as an output thereof and a second source model 32 including an input of the target model as an input thereof, and a combination determination unit configured to combine, when association between an input of the first source model and an output of the second source model is available, the input of the first source model and the output of the second source model.