Data Expansion Algorithm for Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data expansion algorithms often result in unintended data transformations, such as reversing the meaning of images, leading to incorrect classification results, which hinders the achievement of intended learning outcomes in machine learning models.
Innovation Solution
An information processing method that utilizes a processor to acquire expanded data through optional data expansion algorithms, including coupled functions with weights or integral/fractional order operations, and implements learning by stepwise changing these parameters to identify boundary weights or orders that produce intended results, associating them with target data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If optional data expansion algorithms are used to expand training data, then the quantity of training data is increased, but the accuracy of learning results deteriorates due to unintended data transformations
Solution Approach 1:
The patent applies parameter changes by systematically varying transformation parameters (such as rotation angles, scaling factors, translation distances) to identify optimal ranges that preserve data semantics. By adjusting these parameters within specific boundaries, the system generates expanded training data that maintains both quantity increase and accuracy requirements.
Solution Approach 2:
The patent implements dynamics by making the data expansion process adaptive and dynamic. Instead of using fixed transformation rules, the system dynamically adjusts transformation parameters based on feedback from learning results, allowing the expansion algorithm to evolve and optimize its behavior to maintain data quality while increasing quantity.
2Adaptability or versatility
If data expansion algorithms transform images to increase data variety, then the diversity of training data is improved, but the reliability of data meaning deteriorates due to unintended semantic reversals
Solution Approach 1:
The patent applies feedback by using the learning results as a validation mechanism for data expansion. The system monitors whether expanded data maintains correct semantic meaning through learning performance feedback, and adjusts expansion parameters accordingly to prevent semantic reversals while maintaining data diversity.
Solution Approach 2:
The patent implements preliminary action by establishing predetermined boundaries and constraints for data transformation parameters before expansion occurs. By pre-defining safe transformation ranges that preserve semantic meaning, the system prevents harmful transformations while still allowing sufficient diversity in the expanded data.
3Manufacturing precision
If boundary weights and integral orders are determined through systematic learning to identify significant data expansion algorithms, then the manufacturing precision of data expansion is improved, but the loss of time increases due to iterative learning processes
Solution Approach 1:
The patent applies preliminary action by pre-determining boundary weights and integral orders through systematic learning before actual data expansion. By performing this calibration work in advance, the system establishes reusable parameters that can be applied to multiple datasets, reducing the time cost for subsequent expansion operations while maintaining high precision.
Solution Approach 2:
The patent implements partial action by focusing the iterative learning process only on critical parameters (boundary weights and integral orders) rather than optimizing all aspects of data expansion. This selective approach achieves sufficient precision for the most important parameters while minimizing the time investment required.
Data Source
AI summary
It is intended to provide a significant data expansion algorithm for predetermined data. An information processing method performed by a processor included in an information processing device, the method includes: acquiring expanded data resulting from expansion of target data using an optional data expansion algorithm including a coupled function obtained by coupling together a plurality of data expandable functions by using weights; implementing learning, the learning including implementing the learning by inputting the expanded data to a learning model that performs predetermined learning and implementing the learning by using each item of the expanded data generated by stepwise changing a weight of the coupled function; specifying a boundary weight with which a learning result of the learning indicates an intended result and associating the boundary weight with information related to the target data.