Industrial Information Model Selection Using Semantic Confidence Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for mapping human-readable industrial plant documentation into machine-readable formats are time-consuming, expensive, and unreliable due to the variability and domain-specific meanings of the information.
Innovation Solution
A computer-implemented method using trained encoders to transform inputs into embedding spaces for comparison with candidate information models, combined with confidence measures to select the most suitable model or input, ensuring accurate semantic mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If generative machine learning models are used to map documentation to semantic meanings, then processing speed is improved, but reliability deteriorates due to uncontrollable mapping quality
Solution Approach 1:
The patent implements feedback by computing confidence measures for each mapping result and using this information to control the mapping process. The confidence measure provides feedback on the quality of each mapping, allowing the system to adjust its behavior based on reliability metrics rather than operating blindly.
Solution Approach 2:
The patent changes the parameter of confidence threshold dynamically. By adjusting the confidence threshold parameter, the system can balance between processing speed and reliability based on specific requirements. High-confidence mappings are processed quickly while low-confidence ones trigger additional verification or manual review.
2Reliability
If manual perusal of documentation is performed, then mapping reliability is improved, but time consumption and cost increase
Solution Approach 1:
The patent applies partial action by using machine learning models for the majority of mappings that fall within confidence thresholds, while reserving manual review only for edge cases with low confidence scores. This partial automation approach maintains high reliability for critical mappings while reducing overall time consumption compared to fully manual processing.
Solution Approach 2:
The confidence measure acts as an intermediary between automated ML mapping and manual review. It mediates the decision-making process by identifying which mappings require human intervention, thus reducing the burden on manual processors while maintaining reliability where needed.
3Measurement precision
If domain-specific training data is used to improve mapping accuracy, then mapping precision is improved, but data availability and training complexity worsen
Solution Approach 1:
The patent performs preliminary action by pre-computing confidence measures and calibration data during an offline training phase. This preliminary work prepares the model with domain-specific knowledge and confidence calibration, so that during actual operation, the system can make precise mappings without requiring complex real-time computations or extensive domain expert intervention.
Data Source
AI summary
A method for determining an information model includes transforming an input into a representation in an embedding space; comparing this representation to representations of multiple candidate information models in the same embedding space, wherein each information model identifies a collection of related information items with semantic meanings that at least partially characterizes an aspect of the industrial plant; pre-selecting, based on the result of this comparison, from the multiple candidate information models, one or more candidate information models into which the information contained in the given input is likely to fit; computing, for each pre-selected information model, using a given confidence measure, a confidence of the semantic suitability of the respective pre-selected information model for the given input; and selecting an information model with the best confidence of the semantic suitability as the chosen information model into which the information contained in the given input fits.