Model Creation Apparatus for Efficient Transfer Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transfer learning in machine learning is time-consuming and inefficient due to the difficulty in selecting a suitable existing model for new predictions, and managing these models is complex, leading to delayed learning and model selection challenges.
Innovation Solution
A method and apparatus that select a model based on output results from input learning data, create a new model through machine learning, and register it associated with the selected model, facilitating efficient model selection and management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If transfer learning is used to create prediction models by learning a great amount of data, then accurate prediction models can be created in a short time and with a small amount of data, but it takes time and efforts to search for a suitable model among many target models
Solution Approach 1:
The system automatically evaluates existing prediction models by inputting learning data and measuring output performance, using this feedback to rank and select the most suitable model for transfer learning, eliminating manual trial-and-error selection
Solution Approach 2:
The model selection system performs automatic evaluation and ranking of candidate models without requiring manual intervention, allowing the system to self-select the most appropriate existing model for the given learning data and task
2Productivity
If transfer learning is used to create prediction models, then learning time is reduced, but the management of created models becomes complicated and difficult to search
Solution Approach 1:
The system implements a universal model management platform that handles multiple functions including model storage, automatic evaluation, performance ranking, and selection across different prediction tasks, simplifying the management of diverse transferred models through a single integrated system
3Reliability
If manual model selection is performed for transfer learning, then suitable models can be chosen, but learning is delayed due to the time required for selection
Solution Approach 1:
The system performs preliminary automatic evaluation and ranking of candidate prediction models before the actual transfer learning process begins, pre-identifying the most suitable model so that learning can start immediately without selection delays
Data Source
AI summary
A model creation apparatus includes a selector configured to select a model based on output results obtained by inputting pieces of learning data to registered models, a learning unit configured to create a new model by inputting the pieces of learning data to the selected model and performing machine learning, and a registration unit configured to register the created new model such that the new model is associated with the selected model.