Hybrid Data Classification Verification for Edge Computing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data classification methods using neural network models require significant computation and time, especially in edge devices with limited resources, leading to inefficiencies and high costs.
Innovation Solution
A method involving a classifier set with multiple classifiers and a neural network model, where the classifier set determines initial classification, and if it fails to meet a condition, the neural network model is used to verify and provide a classification value, optimizing the classification process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a neural network model is used for data classification, then classification accuracy is improved, but computational cost and time increase significantly
Solution Approach 1:
The patent divides the classification system into multiple classifier sets, each containing multiple classifiers. These classifier sets process different portions or aspects of the input data independently, allowing parallel processing and reducing the overall time required while maintaining accuracy through collective decision-making
2Measurement precision
If a neural network model is used for data classification, then classification accuracy is improved, but computational resources and cost increase
Solution Approach 1:
The classification task is segmented across multiple classifier sets with multiple classifiers each. This distribution allows the computational workload to be divided and executed in parallel, reducing the energy consumption and computational cost compared to a single monolithic neural network while achieving comparable accuracy
Solution Approach 2:
The patent employs multiple classifiers within each classifier set that may use different parameters, thresholds, or processing methods. By varying these parameters across classifiers and aggregating their results, the system achieves high accuracy without requiring a single computationally intensive neural network model
3Measurement precision
If sufficient training data is secured for neural network learning, then classification accuracy is improved, but data requirements and system complexity increase
Solution Approach 1:
Instead of training a single neural network on large volumes of data, the patent segments the learning task across multiple classifiers. Each classifier can be trained on smaller, more manageable datasets, reducing the overall data requirement while the collective output of multiple classifiers achieves high classification accuracy
Data Source
AI summary
Provided is a method for classifying data in an electronic apparatus, including obtaining target data, obtaining first classification information by using a classifier set including a plurality of classifiers based on the target data, obtaining second classification information by using a neural network model based on the target data, comparing the first classification information and the second classification information, and verifying the classifier set based on a result of comparing the first classification information and the second classification information.