Learning Sample Error Layout for Noisy Data Exclusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional learning models suffer from decreased estimation accuracy due to the inclusion of noisy data in the learning samples, which existing methods fail to efficiently identify and mitigate.
Innovation Solution
A learning apparatus that iteratively learns a neural network model, displays learning samples arranged based on their errors, and allows users to set boundary lines to exclude noisy data, thereby improving estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If learning is performed using all learning samples including noisy data, then the learning process can be completed with all available data, but the estimation accuracy of the learning model decreases
Solution Approach 1:
The patent extracts and removes noisy data from the learning samples through automated detection mechanisms. The system identifies samples with high learning errors or inconsistent patterns and excludes them from the learning process, thereby improving estimation accuracy while maintaining the use of the majority of valid learning data.
Solution Approach 2:
The patent changes the parameters of learning samples by dynamically adjusting which samples are included in the learning process. Through automated noise detection and sample selection, the system modifies the composition of learning data based on detected error patterns and sample quality metrics.
2Reliability
If noisy data is eliminated from learning samples, then the estimation accuracy improves, but the complexity of identifying and removing noisy data increases
Solution Approach 1:
The patent implements a self-service mechanism where the learning apparatus automatically detects and identifies noisy data through built-in error detection algorithms. The system monitors learning errors, compares sample patterns, and autonomously determines which samples are noisy without requiring external manual intervention or complex external analysis systems.
Solution Approach 2:
The patent employs feedback mechanisms by continuously monitoring learning errors and using this information to identify noisy samples. The system feeds back error patterns to the sample selection process, adjusting which samples are included in learning based on detected performance issues, thereby simplifying noise detection through iterative self-correction.
3Stability of the object's composition
If learning is performed with noisy data, then all available data is utilized, but the learning stability decreases
Solution Approach 1:
The patent applies preliminary action by detecting and removing noisy data before the learning process begins. The system performs preliminary analysis of learning samples to identify problematic data, excluding these samples from the learning process in advance, thereby ensuring learning stability from the outset while maintaining a substantial dataset.
Data Source
AI summary
According to one embodiment, a learning apparatus includes processing circuitry. The processing circuitry acquires a plurality of learning samples to be learned and a plurality of target labels associated with the respective learning samples, iteratively learns a learning model so that a learning error between output data corresponding to the learning sample and the target label is small with respect to the learning model to which the output data is output by inputting the learning sample, and displays a layout image in which at least some of the learning samples are arranged based on a learning progress regarding the iterative learning of the learning model and a plurality of the learning errors.