Geolocation Estimation Model Self-Evaluation via Quality Metrics

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current geolocation estimation models rely heavily on independent test data, which is costly and limited in representing full spatial and temporal variations, leading to inaccurate geolocation estimates and failure to detect edge cases, resulting in inefficient resource utilization and poor accuracy profiles.

Innovation Solution

A method that utilizes quality evaluator models to process geolocation input data and estimates to generate quality evaluation metrics, allowing for the modification of geolocation estimation models without independent test data, thereby improving estimate consistency and plausibility.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If independent test data is used to evaluate geolocation estimation models, then measurement precision is improved, but loss of information increases due to inability to represent full spatial and temporal variations

Engineering Contradiction:
Improvegeolocation estimate accuracyVSAvoidspatial and temporal variation coverage
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The geolocation estimation model performs self-evaluation by processing its own output estimates through quality evaluator models. The model uses its own predictions and associated input data to generate quality evaluation metrics, eliminating the need for external independent test data and enabling comprehensive coverage of all spatial and temporal variations in the service area.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

Quality evaluator models serve as intermediaries between the geolocation estimation model and the evaluation process. These evaluator models process the geolocation estimates and input data to generate quality metrics, acting as a mediator that enables self-evaluation without requiring independent test data while maintaining assessment capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If extensive independent test data is collected to cover all edge cases, then reliability is improved, but loss of time increases due to data collection and processing requirements

Engineering Contradiction:
Improveedge case detection capabilityVSAvoiddata collection and processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs self-evaluation by processing existing operational data through quality evaluator models. Instead of collecting extensive independent test data covering all edge cases, the model uses its own predictions and associated input data to identify reliability issues and edge cases, dramatically reducing data collection and processing time while maintaining comprehensive evaluation capability.

Inventive Principle:
Principle #25Self-service

3Manufacturing precision

If quality evaluator models process multiple sets of geolocation input data to generate comprehensive quality metrics, then manufacturing precision is improved, but use of energy increases

Engineering Contradiction:
Improvemodel performance accuracyVSAvoidcomputing resource consumption
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The geolocation estimation model performs self-evaluation using its own output and input data, eliminating the need for separate independent test datasets. This approach generates comprehensive quality metrics through processing existing data multiple times with different quality evaluator models, improving model precision while avoiding the energy cost of collecting and processing extensive external test data.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4224116A1Utilizing models to evaluate geolocation estimate quality without independent test data
Publication Date: 2023.08.09 VIAVI SOLUTIONS INC(US)
  • EP4224116A1 patent drawingFigure 1A
  • EP4224116A1 patent drawingFigure 1B
  • EP4224116A1 patent drawingFigure 1C

AI summary

A device may receive a geolocation estimation model and geolocation input data identifying geolocations. The device may process one or more sets of the geolocation input data, with a quality evaluator model, to generate one or more quality evaluation metrics. The device may modify the geolocation estimation model based on the one or more quality evaluation metrics.