Industrial Monitoring Data Selection for AI Target State Labeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The inefficiency in identifying and annotating relevant data subsets from large datasets in industrial processes for machine learning model training and evaluation, due to labor-intensive human annotation and subjective quality variations, poses a challenge in automating industrial practices like defect detection and process control.
Innovation Solution
A computer-implemented method that identifies target states in industrial processes using monitoring data from various sources, generating target data based on these states, which can be used to improve machine learning models, employing state detection models like neural networks to automate data selection and reduce human intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human annotators manually analyze and annotate collected data to identify relevant subsets, then data quality and relevance can be ensured, but the process becomes very time-consuming and labor-intensive
Solution Approach 1:
The patent introduces an automated data selection system that acts as an intermediary between raw monitoring data and human annotators. The system uses machine learning models and algorithms to pre-process and filter large datasets, identifying and selecting relevant data subsets before presenting them to human annotators. This intermediary layer significantly reduces the time and effort required for manual annotation while maintaining data quality through automated objective criteria.
Solution Approach 2:
The patent replaces the mechanical process of manual data analysis and annotation with automated computational methods. Machine learning models, pattern recognition algorithms, and automated selection criteria substitute for human annotators in identifying relevant data subsets. This substitution dramatically reduces annotation time and labor requirements while maintaining or improving consistency and objectivity in data selection.
2Adaptability or versatility
If human annotators subjectively identify relevant data subsets, then flexibility in judgment can be applied, but quality and relevance vary due to human subjectivity
Solution Approach 1:
The patent implements feedback mechanisms where automated selection results are continuously refined based on performance metrics and validation. The system uses objective criteria and algorithms that provide consistent, repeatable data selection across different datasets and annotators. Feedback loops allow the system to learn from results and improve selection accuracy while maintaining reliability through standardized automated processes rather than subjective human judgment.
Solution Approach 2:
The patent transforms the subjective parameter of human judgment into objective computational parameters. Instead of relying on annotator subjectivity, the system uses defined parameters, algorithms, and mathematical criteria to identify relevant data subsets. This parameter transformation ensures consistent and reliable data selection that can be replicated and validated across different contexts while maintaining the flexibility to adjust selection criteria as needed.
3Loss of information
If all collected monitoring data is processed and stored, then complete information is available, but data handling becomes cumbersome with high volume requirements
Solution Approach 1:
The patent extracts only the essential and relevant features, patterns, and data subsets from large volumes of monitoring data using automated selection algorithms. Instead of processing and storing all raw data, the system identifies and extracts the critical information needed for machine learning model training and evaluation. This extraction approach maintains data completeness for relevant information while dramatically reducing the volume of data that needs to be handled, stored, and processed.
Solution Approach 2:
The patent segments large volumes of monitoring data into meaningful subsets based on relevance, quality, and utility for specific machine learning tasks. The automated selection system divides the comprehensive dataset into targeted segments that are most valuable for training and evaluation. This segmentation maintains the completeness of essential information while organizing data into manageable segments that reduce handling complexity and improve processing efficiency.
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3B
AI summary
A computer implemented method for determining or generating target data for artificial intelligence is disclosed. In one aspect, the method may include identifying an indication of a target state of a first industrial process or asset, or of a second industrial process or asset operatively associated with the first industrial process or asset, and determining or generating target data, based on the identified indication of the target state. The identification of the indication of the target state may be based on first monitoring data indicative of data output from first monitoring data source(s) associated with the first industrial process or asset or with the second industrial process or asset. The target data may be determined or generated from second monitoring data indicative of data output from second monitoring data source(s) associated with the first industrial process or asset.