Sample Clustering for Learning Data Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high data acquisition rates from sensors result in enormous sample sizes, leading to communication band constraints and power consumption issues when transmitting all samples to a learning apparatus.
Innovation Solution
A method where samples are divided into clusters, and only the most effective samples are transmitted to the learning apparatus, with the effectiveness of each cluster calculated based on learning results to optimize data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all samples are transmitted to the learning apparatus, then learning accuracy is improved, but communication bandwidth and power consumption increase
Solution Approach 1:
The patent segments the enormous number of samples into manageable clusters using clustering algorithms. By dividing the data into clusters and selecting representative samples from each cluster, the system reduces the total number of samples to be transmitted while preserving the essential information needed for accurate learning.
Solution Approach 2:
The patent extracts only the most effective and representative samples from each cluster for transmission to the learning apparatus. This extraction is guided by effectiveness metrics that evaluate which samples contribute most to learning accuracy, thereby transmitting only the necessary data and reducing overall transmission volume.
2Measurement precision
If all samples are transmitted to the learning apparatus, then learning accuracy is improved, but power consumption of devices increases
Solution Approach 1:
The patent segments the data transmission process into clustering at the device level followed by selective sample extraction and transmission. This segmentation allows the device to process and filter data locally before transmission, significantly reducing the amount of data that consumes communication energy and device power.
Solution Approach 2:
The patent extracts only the most effective samples from each cluster for transmission, guided by effectiveness calculations. This extraction minimizes the number of samples transmitted, thereby reducing the power consumption associated with data transmission while maintaining learning accuracy.
3Quantity of substance
If samples are selectively extracted and transmitted, then communication bandwidth and power consumption are reduced, but learning accuracy may deteriorate
Solution Approach 1:
The patent implements a feedback mechanism where the learning apparatus evaluates the effectiveness of transmitted samples and provides this information back to the device. This feedback loop enables the system to iteratively improve sample selection, ensuring that the extracted samples remain representative and effective for maintaining high learning accuracy despite the reduced data volume.
Solution Approach 2:
The patent replaces the mechanical approach of transmitting all raw data with an intelligent selection mechanism based on effectiveness metrics. By substituting brute-force data transmission with a smart sampling strategy guided by learning results, the system maintains accuracy while reducing data transmission requirements.
Data Source
AI summary
A learning method includes a step in which a device acquires a plurality of samples, a step in which the device divides the plurality of samples into a plurality of clusters, a step in which the device extracts samples from each of the plurality of clusters according to an effectiveness of each cluster received from a learning apparatus, a step in which the device transmits the extracted samples to the learning apparatus, a step in which the learning apparatus learns the extracted samples, a step in which the learning apparatus calculates, for each cluster, an effectiveness in learning the samples belonging to a cluster from learning results, and a step in which the learning apparatus transmits the effectiveness of each cluster to the device.