Machine Learning Algorithm Composition via Pre-generated Learning Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The machine learning system faces challenges in efficiently composing image processing algorithms with high accuracy due to the need for proper skill and sufficient learning patterns, making it difficult to utilize its characteristics effectively.
Innovation Solution
An information processing apparatus and method that creates multiple learning information items, including input and teacher images, using a program code scenario, and supplies them to a machine learning module to compose an image processing algorithm, with the ability to select basic functions and debug control signals to optimize the learning process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning is used to compose image processing algorithms, then algorithm composition becomes automated, but proper skill and learning pattern selection are still required which reduces ease of operation
Solution Approach 1:
The system pre-generates multiple learning patterns and stores them in a database before the actual algorithm composition process. When a user wants to compose an algorithm, the system automatically selects and applies pre-prepared learning patterns without requiring the user to manually configure learning parameters, thus achieving both automation and ease of operation
Solution Approach 2:
The system automatically evaluates multiple learning patterns against the target algorithm composition task and selects the most appropriate one without external intervention. The system self-manages the entire process from pattern selection to algorithm generation, reducing the need for user expertise while maintaining high automation
2Manufacturing precision
If multiple learning information items are generated to improve learning accuracy, then the quality of the composed algorithm improves, but the time and computational resources required increase
Solution Approach 1:
The system pre-generates a large number of diverse learning information items and stores them in a database before the actual algorithm composition. This allows the system to quickly retrieve and apply appropriate learning patterns during composition without spending excessive time generating data on-demand, thus improving precision while controlling time consumption
Solution Approach 2:
The system generates more learning information items than strictly necessary (excessive action) and stores them in advance. During algorithm composition, it selectively uses only the portion needed, achieving high precision through the availability of abundant pre-generated data without paying the full time cost of generating all data during the composition process
Data Source
AI summary
An information processing apparatus includes a control unit that creates a plurality of learning information items including an input image and a teacher image as an expected value by image-processing the input image in accordance with a scenario described with a program code, and supplies the created plurality of learning information items to a machine learning module that composes an image processing algorithm by machine learning.