Neural Network Connection Segmentation for Classification Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pattern classification methods using artificial neural networks face challenges in efficiently updating connection intensities to accurately classify input patterns, particularly in applications like medical diagnostics where precise classification of biological signals is required.
Innovation Solution
The method involves enabling an artificial neural network to learn based on input data, updating connection intensities between hidden and output layers using Principal Component Analysis (PCA) and Extreme Learning Machine (ELM) methods, allowing for effective classification of input patterns by identifying and optimizing connection intensities for improved pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If connection intensities of all connection lines are updated during learning, then classification accuracy may improve, but learning time and computational complexity increase significantly
Solution Approach 1:
The patent segments the connection lines into two distinct groups: those connecting the input layer to the hidden layer, and those connecting the hidden layer to the output layer. During the learning process, only the connection intensities of the second group (hidden layer to output layer) are updated, while the first group remains fixed. This segmentation allows the system to maintain classification accuracy while significantly reducing the computational burden and learning time associated with updating all connection intensities.
2Measurement precision
If more connection lines are updated to improve pattern recognition accuracy, then classification precision improves, but device complexity and computational resources required increase
Solution Approach 1:
The patent divides the neural network's connection lines into two functional segments: input-to-hidden connections that remain fixed, and hidden-to-output connections that are updated during learning. This segmentation strategy maintains pattern recognition accuracy by preserving the feature extraction capability of the hidden layer while reducing computational complexity by limiting updates to only the necessary connection intensities.
3Adaptability or versatility
If traditional learning methods update all connection intensities, then the neural network can adapt to various patterns, but the learning process becomes inefficient and time-consuming
Solution Approach 1:
The patent segments the learning process into two distinct phases: a first learning phase where input-to-hidden connection intensities are established and fixed, and a second learning phase where only hidden-to-output connection intensities are updated. This segmented approach maintains the neural network's adaptability to various patterns while dramatically improving learning efficiency by avoiding redundant updates to connection intensities that do not contribute to pattern classification.
Data Source
AI summary
A method of classifying an input pattern and a pattern classification apparatus are provided. The method includes enabling an artificial neural network to learn based on learning input data received by an input layer of the artificial neural network, determining classification of an input pattern received by the input layer of the enabled artificial neural network according to an output value obtained from an output layer of the artificial neural network, the obtained output value being based on the input pattern, updating connection intensities of a plurality of connection lines of the enabled artificial neural network to output a result value indicating the determined classification from the output layer when the input pattern, and determining updated classification of the input pattern according to an updated output value obtained from an output layer of the updated artificial neural network, the obtained updated output value being based on the input pattern.