Regression Processing Using Vector Neural Network Feature Spectra
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a lack of effective techniques for highly accurate regression processing using vector neural networks, as existing applications have not sufficiently explored their potential in this area.
Innovation Solution
A regression processing device and method utilizing a machine learning model with multiple vector neuron layers, which includes a regression processing unit that calculates a degree of similarity between a known feature spectrum group and a feature spectrum obtained from an output of a specific layer to predict output values, enabling high-accuracy regression processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If vector neural network is applied to regression processing, then the model structure can be simplified, but the processing accuracy is insufficient
Solution Approach 1:
The patent applies preliminary action by pre-extracting and storing feature spectra from teaching data before actual regression processing. The memory stores known feature spectrum groups obtained from teaching data, so that during new data processing, the system can directly compare and match against these pre-computed features, improving accuracy without adding complex real-time processing steps
Solution Approach 2:
The patent introduces feature spectrum as an intermediary between the input data and the regression output. Instead of directly mapping input data to output values, the system extracts feature spectra from both teaching and new data, compares them, and uses the degree of similarity to determine regression results, thereby bridging the gap between simple model structure and accurate prediction
2Measurement precision
If feature spectrum comparison is added to improve accuracy, then regression processing accuracy is improved, but processing time increases
Solution Approach 1:
The system performs feature spectrum extraction and storage in advance during the training phase with teaching data. This preliminary action eliminates the need to re-extract and re-process features during new data prediction, significantly reducing processing time while maintaining high accuracy through the pre-computed feature comparisons
Solution Approach 2:
The patent creates a copy of the teaching data feature spectra and stores them in memory for later retrieval. During new data processing, instead of re-computing everything from scratch, the system retrieves the pre-computed feature spectrum copies and compares them with new data features, reducing computational load and processing time while preserving accuracy
Data Source
AI summary
A regression processing unit is configured to execute processing (a) of obtaining a predicted output value with respect to input data using a machine learning model, processing (b) of reading out a known feature spectrum group from a memory, processing (c) of calculating a degree of similarity relating to the predicted output value between the known feature spectrum group and a feature spectrum obtained from an output of a specific layer when the input data is input to the machine learning model, and processing (d) of outputting the predicted output value using the degree of similarity.