Measurement Data Event Detection With Modular ML Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for event detection in measurement data, such as OTDR, require extensive training data and are prone to overfitting, necessitating expert evaluation and being inflexible to various conditions.
Innovation Solution
A method utilizing separate feature extraction and detection models, trained with limited data and optimized through user feedback, enabling scalable and adaptable event detection across diverse scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex machine learning algorithms are used for event detection, then detection accuracy is improved, but the requirement for training data and computational resources increases
Solution Approach 1:
The patent divides the complex machine learning algorithm into separate sub-models that are implemented and trained independently. This segmentation allows each sub-model to be trained on smaller, more manageable datasets while maintaining overall detection accuracy, thereby reducing the total training data requirement.
Solution Approach 2:
The patent introduces intermediate representations or features that serve as mediators between the input data and final event detection. These intermediate layers break down the complex detection task into simpler stages, each requiring less training data while collectively achieving high detection accuracy.
2Measurement precision
If complex machine learning algorithms are used for event detection, then detection accuracy is improved, but computational resources and storage space increase
Solution Approach 1:
By segmenting the complex algorithm into separate sub-models, the computational load is distributed across multiple smaller models rather than one large model. This reduces the memory and computational resources required at any given time while maintaining detection accuracy through the coordinated operation of sub-models.
Solution Approach 2:
The patent uses simpler sub-models that replicate or approximate the functionality of a single complex model. These simpler copies require fewer computational resources and storage space while collectively achieving the same detection performance when combined.
3Adaptability or versatility
If a single comprehensive model is used, then all patterns can be detected, but the model becomes inflexible and difficult to adapt to different conditions
Solution Approach 1:
The patent segments the detection system into multiple independent sub-models, each potentially specialized for different aspects of event detection. This segmentation increases flexibility as individual sub-models can be developed, trained, and optimized separately, making the overall system more adaptable to different conditions while managing complexity through modular architecture.
Solution Approach 2:
The patent creates a dynamic system where separate sub-models can be selectively activated or combined based on the specific detection task or conditions. This dynamic configuration allows the system to adapt its complexity and capabilities to match the requirements of different scenarios, improving versatility while avoiding unnecessary complexity.
Data Source
AI summary
The present invention relates to a method for detecting an event based on data acquired by a measurement device, wherein the method (1) comprises the steps of:Acquiring data by a measurement device (2);Extracting features from the acquired data utilizing a feature extraction model (3);Detecting an event based on the extracted features utilizing a detection model (4); andProviding information about the detected event (5).