Feature Transformation Device for Domain Adaptation Precision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning techniques face precision issues due to differences in statistical characteristics between training and test data, with current domain adaptation methods either reducing the number of effective sample data or using unrelated training data for feature transformation.
Innovation Solution
A feature transformation device and method that optimizes weight and feature transformation parameters using an objective function with regularization, ensuring that the statistical characteristics of training and test data are approximated while increasing the number of effective sample data by weighting and transforming both datasets effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature transformation is performed using existing domain adaptation techniques, then statistical characteristics of training and test data are approximated, but precision of learning is affected due to using unrelated training data
Solution Approach 1:
The patent applies local quality by assigning different weights to different training data samples based on their relevance to the target domain. The weight derivation unit calculates specific weights for each training sample, allowing the system to selectively emphasize relevant data while downweighting unrelated data, thereby improving both domain adaptation precision and learning reliability
Solution Approach 2:
The patent changes parameters by introducing weight parameters and regularization parameters that control the feature transformation process. By optimizing these parameters through the objective function with regularization terms, the system adapts the transformation to maximize precision while maintaining learning reliability
2Measurement precision
If weight is assigned to each element in training data without regularization, then effective sample data may be decreased, but optimization precision may be improved
Solution Approach 1:
The patent applies beforehand cushioning by incorporating regularization terms into the objective function before optimization. The regularization unit adds penalty terms that prevent excessive weight values, cushioning against the risk of reducing effective sample data while still allowing precise optimization of relevant features
Solution Approach 2:
The patent implements feedback through the iterative optimization process where the objective function evaluation provides feedback on weight assignments. The system adjusts weights based on this feedback while regularization ensures that adjustments don't excessively reduce the number of effective samples, balancing precision and quantity
Data Source
AI summary
Provided are a feature transformation device and others enabling feature transformation with high precision.The feature transformation device includes receiving means for receiving training data and test data each including a plurality of samples, optimization means for optimizing weight and feature transformation parameter based on an objective function related to the weight and the feature transformation parameter, the optimization means including weight derivation means for deriving the weight assigned to each element included in the training data and feature transformation parameter derivation means for deriving the feature transformation parameter that transforms each of the samples included in the training data or the test data, objective function derivation means for deriving a value of the objective function, the objective function derivation means including a constraint determination means for determining whether the weight satisfies a prescribed constraint and regularization means for regularizing at least one of the weight or the feature transformation parameter, and transformation means for transforming an element included in at least one of the training data or the test data based on the feature transformation parameter.