Input Data Selection by Prediction Quality Trend Matching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing machine-learned models face challenges such as insufficient data availability, lengthy development processes, insufficient reliability, and inability to handle uncertainties, particularly in complex environments, leading to limited applicability and high implementation barriers.

Innovation Solution

A method and system for selecting input data points using a machine-learned prediction model that involves training the model in multiple steps, determining prediction quality values, and verifying similarity to reference trends to ensure high processing quality, allowing for reliable and automated data processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If machine-learned models are used for automated processing, then processing automation and efficiency are improved, but data availability and quality requirements worsen

Engineering Contradiction:
Improveprocessing automationVSAvoiddata availability
Core Design Contradiction:
Extent of automationVSQuantity of substance

Solution Approach 1:

The patent applies preliminary action by selecting and preparing suitable input data points before they are processed by the machine-learned model. The system proactively identifies and selects data points that are likely to yield reliable predictions, preparing them in advance for processing. This prevents the need for extensive post-processing corrections and ensures that the model receives high-quality input data, thereby reducing the overall data quality burden while maintaining high automation levels.

Inventive Principle:
Principle #10Preliminary action

2Extent of automation

If machine-learned models are used for automated processing, then processing automation is improved, but processing reliability worsens

Engineering Contradiction:
Improveprocessing automationVSAvoidprocessing reliability
Core Design Contradiction:
Extent of automationVSReliability

Solution Approach 1:

The patent implements feedback mechanisms where the system continuously monitors and evaluates the performance of the machine-learned model on selected data points. The system uses this feedback to refine its data selection criteria and improve future selections. This iterative feedback loop ensures that only data points with high prediction quality are processed, thereby maintaining high reliability in automated processing while scaling automation capabilities.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary selection and validation of input data points before processing, ensuring that only high-quality data enters the machine-learned model. This preliminary action filters out potentially unreliable data points in advance, preventing them from affecting processing reliability while maintaining high automation levels.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If complex machine-learned models are used, then processing capability is improved, but development time and complexity worsen

Engineering Contradiction:
Improveprocessing capabilityVSAvoiddevelopment time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-selecting and pre-validating suitable input data points before processing by the complex machine-learned model. This preliminary preparation reduces the effective development time needed for model training and optimization, as the data is already curated and ready for processing. The system can thus leverage complex models more efficiently without proportionally increasing development time.

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If data selection is performed manually, then data quality is improved, but processing time and cost worsen

Engineering Contradiction:
Improvedata qualityVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements self-service by enabling the system to automatically select and evaluate its own input data points using integrated selection criteria and machine-learned evaluation models. The system autonomously identifies suitable data points without requiring manual intervention, thereby maintaining high data quality standards while significantly reducing the time and resources that would otherwise be required for manual data selection and curation.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4679339A1Methods and systems for selecting and processing input data points for processing by a machine-learned prediction model and computer program product
Publication Date: 2026.01.14 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP4679339A1 patent drawingFigure 1
  • EP4679339A1 patent drawingFigure 2
  • EP4679339A1 patent drawingFigure 3

AI summary

The invention relates to a method and system for selecting input data points for processing by a machine-learned prediction model, and to a method for processing input data points by a machine-learned prediction model generated by machine learning in several training steps (TS), comprising the steps of: - acquiring or receiving an input data point (EDP), - determining a trend of at least two prediction quality values ​​(PQW), wherein these prediction quality values ​​(PQW) are each determined with a training-step-specific instance of a selection model, - checking whether the trend of the at least two prediction quality values ​​(PQW) is similar to at least one reference trend (R1, R2, ..., Rm), - selecting the input data point (EDP) for processing if the trend of the at least two prediction quality values ​​(PQW) is similar to the at least one reference trend (R1, R2, ..., Rm).