Learning Device Signal Clustering Termination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for learning signal waveforms often end learning at an inappropriate time, leading to inefficiencies in abnormality detection in production systems, as they do not accurately determine when sufficient pattern convergence is achieved.
Innovation Solution
A learning device that acquires and clusters partial signals from the learning signal based on similarity, generating progress information to determine the appropriate termination point of learning, ensuring that learning is stopped when the clustering process reaches a certain degree of similarity and stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If learning continues after convergence of the number of patterns, then learning accuracy may be improved, but calculation resources are wasted
Solution Approach 1:
The system monitors the number of patterns during learning and uses this feedback to determine when to terminate learning. When the number of patterns converges (stops increasing significantly), the system automatically terminates learning, preventing waste of calculation resources while ensuring sufficient learning accuracy is achieved.
Solution Approach 2:
The learning termination condition is dynamically adjusted based on the convergence behavior of the number of patterns. Rather than using a fixed termination criterion, the system adapts the termination point based on how the pattern count evolves during learning, allowing optimal balance between accuracy and resource consumption.
2Loss of energy
If learning terminates based on convergence of the number of patterns, then calculation resources are saved, but learning may end at an inappropriate time for waveform comparison
Solution Approach 1:
The system transitions from monitoring only the number of patterns to also monitoring the distribution of similarities between waveforms. This additional dimension of evaluation ensures that learning terminates not just when pattern count stabilizes, but also when waveform similarity distribution indicates sufficient learning for accurate abnormality detection.
Solution Approach 2:
The termination criterion incorporates changes in the distribution of similarities as an additional parameter alongside the number of patterns. By monitoring how similarity distribution evolves and stabilizes, the system ensures learning terminates at an appropriate point that guarantees both resource efficiency and detection reliability.
3Measurement precision
If learning of waveforms proceeds after convergence of pattern number, then waveform comparison accuracy may improve, but learning time increases
Solution Approach 1:
The system uses feedback from the distribution of similarities to determine when waveform learning has reached sufficient accuracy. When the similarity distribution stabilizes within a certain range, the system terminates learning, preventing unnecessary extension of learning time while ensuring adequate waveform comparison accuracy is achieved.
Data Source
AI summary
A learning device (10) includes an acquirer (11), a learner (12), and a generator (14). The acquirer (11) acquires a learning signal. The learner (12) performs, in accordance with similarities indicating degrees of similarity between waveforms, clustering of partial signals cut out from the learning signal acquired by the acquirer (11), and learns reference waveforms that each indicate a waveform of a corresponding partial signal of the clustered partial signals. The generator (14) generates, based on at least one of a distribution of the similarities or characteristics of clusters that each include a corresponding partial signal of the clustered partial signals, progress information indicating a progress status of the learning by the learner (12), and outputs the progress information.