Deep Learning Neural Records Incrementally Refined Through Expert Input
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Rule-based classification systems are inefficient for large, varying, and complex data sets, prone to failure with changing data, and difficult to maintain, especially when manual design becomes complex and feature identification for automatic classification is challenging.
Innovation Solution
A deep learning neural network model is used for data classification, incrementally refined through expert input, where real-time network information is classified, and if errors exceed thresholds, classifiers are tuned, and a fast learning model is trained to adapt, with confidence values comparing deep learning and fast learning model classifications to determine the most accurate results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If rule-based classification systems are used, then classification can be performed with simple data sets, but the systems become difficult and expensive to maintain with large, varying, and complex data sets
Solution Approach 1:
The patent replaces manual rule-based classification systems with automated machine learning models that learn classification patterns from data. This substitution eliminates the need for manual rule creation and maintenance while providing adaptability to complex and varying data sets through automatic learning and refinement processes.
Solution Approach 2:
The machine learning models perform self-service by automatically learning from training data, identifying patterns, and improving their own classification capabilities without requiring manual intervention. The system autonomously refines its performance through continuous learning from new data while maintaining simplicity for the user.
2Ease of operation
If manual rule-based classifiers are designed, then classification can be performed, but the process becomes difficult as classification options become more complex
Solution Approach 1:
The patent replaces the manual mechanical process of designing classification rules with automated machine learning algorithms. These algorithms automatically generate classification logic based on training data, eliminating the difficulty of manually designing complex classifiers while handling sophisticated classification options through automatic feature extraction and pattern recognition.
3Measurement precision
If deep learning neural network is used for classification, then accuracy improves on complex data sets, but training time and computational resources increase
Solution Approach 1:
The patent performs preliminary action by pre-training deep learning models on large data sets before deployment. This advance training prepares the models to quickly adapt to specific classification tasks with minimal additional training time, reducing the loss of time during actual operational use while maintaining high accuracy on complex data sets.
Solution Approach 2:
The patent implements dynamic training approaches where the model adapts its training intensity and duration based on the specific task requirements. The system can perform quick fine-tuning on new data without requiring complete retraining, allowing the model to maintain high accuracy while minimizing training time through adaptive learning rates and selective retraining on relevant data subsets.
Data Source
AI summary
Embodiments are directed towards classifying data using machine learning that may be incrementally refined based on expert input. Data provided to a deep learning model that may be trained based on a plurality of classifiers and sets of training data and/or testing data. If the number of classification errors exceeds a defined threshold classifiers may be modified based on data corresponding to observed classification errors. A fast learning model may be trained based on the modified classifiers, the data, and the data corresponding to the observed classification errors. And, another confidence value may be generated and associated with the classification of the data by the fast learning model. Report information may be generated based on a comparison result of the confidence value associated with the fast learning model and the confidence value associated with the deep learning model.