Joint Auto-Encoder and Classifier Training for Medical Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical data, characterized by high noise, multivariate high-dimensionality, and heterogeneity, poses challenges in classification due to error transmission and accumulation during independent training of auto-encoding and classification models, leading to reduced accuracy and increased training time.
Innovation Solution
A method involving the joint training of an auto-encoding module and a classification module using medical sample data, where the auto-encoding module processes data to obtain reconstructed sample data, and the classification module processes feature data to generate classification results, with supervised labels introduced to mitigate unsupervised auto-encoding issues and error accumulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If auto-encoding module and classification module are trained independently, then training flexibility is maintained, but error transmission and accumulation occur leading to reduced accuracy
Solution Approach 1:
The patent merges the auto-encoding module and classification module into a unified joint training framework. The encoder extracts features while the classifier performs classification, and both are optimized simultaneously through a combined loss function that includes reconstruction loss and classification loss. This integration eliminates error transmission between independently trained modules while maintaining manageable complexity through modular architecture design.
2Reliability
If traditional independent training methods are used, then training time is reduced, but model accuracy deteriorates due to error accumulation
Solution Approach 1:
The joint training framework enables continuous optimization of both encoding and classification functions simultaneously. The backpropagation algorithm continuously adjusts parameters of both modules in an integrated manner, ensuring that feature extraction and classification are optimized together without interruption or error accumulation, thereby achieving higher accuracy efficiently.
3Measurement precision
If multiple model iterations are performed to improve accuracy, then classification precision increases, but training time and computational burden increase
Solution Approach 1:
The encoder in the joint training framework performs preliminary feature extraction and dimensionality reduction before the classification step. By pre-processing the data through the encoder to extract essential features, the classification module receives refined input that requires fewer iterations to achieve high precision, thereby reducing overall training time and computational burden.
Data Source
AI summary
A method and an apparatus of training a classification model, a classification method, a classification apparatus, an electronic device, and a medium are provided. The method includes: processing first sample data by using an auto-encoding module, so as to obtain reconstructed sample data, wherein the auto-encoding module includes at least one autoencoder, the autoencoder includes an encoder and a decoder, and the first sample data includes medical sample data; processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result; jointly training the auto-encoding module and the classification module according to the first sample data, the reconstructed sample data, the first sample classification result, and a first sample classification label value of the first sample data; and obtaining the classification model according to the trained encoder and the trained classification module.