Machine-Learning Plausibility Checks for Technical Operating States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous systems, such as autonomous trains and industrial plants, lack a reliable method to ensure accurate classification of obstacles and personnel detection, which can lead to dangerous situations due to faulty machine learning models.
Innovation Solution
A computer-implemented method using multiple machine learning approaches to determine the operating state of data elements and assign plausibility indicators, incorporating expert knowledge of the operational domain to validate the accuracy of classifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single machine learning approach is used for classification, then the system complexity is low and processing is fast, but the reliability and accuracy of classification results cannot be guaranteed
Solution Approach 1:
The patent divides the classification task into multiple independent machine learning approaches (first approach and second approach), each processing the input dataset separately to produce independent classification results and plausibility indicators. This segmentation allows the system to compare results across different approaches, thereby improving reliability without requiring a single overly complex model
Solution Approach 2:
The patent creates multiple copies of the classification process using different machine learning approaches. Each approach independently classifies the same input data, generating duplicate but independent plausibility indicators. By comparing these copied classification results, the system can validate accuracy and improve reliability while maintaining manageable complexity in each individual approach
2Reliability
If multiple machine learning approaches are used to improve classification reliability, then the accuracy and confidence estimation improve, but the computational time and processing speed increase
Solution Approach 1:
The patent performs preliminary classification using the first machine learning approach to generate initial plausibility indicators quickly. This first pass provides immediate classification results that can be used for time-critical decisions, while the second approach runs in parallel or subsequently to provide additional validation. This preliminary action reduces the perceived processing time for critical operations while still benefiting from multiple approaches
3Measurement precision
If different input datasets are used for the first and second approaches, then the validation of classification accuracy is improved, but the data processing complexity and resource requirements increase
Solution Approach 1:
The patent applies different input datasets strategically to different machine learning approaches based on their specific strengths and requirements. The first approach may use a dataset optimized for its particular algorithm, while the second approach uses a different dataset that validates specific aspects of classification. This local quality approach allows validation of accuracy without requiring complete uniformity across all data processing paths, thereby managing complexity
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a computer-implemented method for checking the plausibility of data elements, comprising the steps: a. providing a first input data set comprising a plurality of data elements (S1); b. providing a second input data set comprising a plurality of data elements (S2); wherein the input data sets are either different from or the same; c. determining an associated operating state for each data element of the plurality of data elements by means of a first approach based on machine learning based on the first input data set or the second input data set (S3); wherein the operating state relates to a technical system; d.Determining an associated operating state for each data element of the majority of the data elements by means of at least one second approach based on machine learning based on the first input data set or the second input data set (S4); e. Assigning a plausibility marker to each data element that is classified by means of the first approach and/or assigning a plausibility marker to each data element that is classified by means of the second approach (S5); and f. Providing the data elements classified by means of the first approach with their respective associated plausibility markers and providing the data elements classified by means of the second approach with their respective associated plausibility markers (S6). The invention further relates to a technical system and a corresponding computer program product.