Inference Unit Parameter Variation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the performance of structural complexes using neural networks only provide a single estimation value, making it difficult for users to understand how individual parameters influence the predicted performance, leading to reduced user convenience and inference processing efficiency as the number of parameters increases.
Innovation Solution
A data processing apparatus and method that uses a trained model to predict objective variables from explanatory variables, allowing for the variation of a selected explanatory variable within a predetermined range while keeping others fixed, and generates data to display the variation of objective variables, enhancing the understanding of parameter influence on the inference result.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single estimation value is provided for a data set of multiple parameters, then the inference processing is simple and fast, but the user cannot understand how each parameter influences the predicted performance
Solution Approach 1:
The patent segments the single estimation value into multiple individual parameter influence values by performing separate inference processing for each parameter. This allows the system to provide detailed information about how each parameter affects the predicted performance while maintaining efficient processing through automated sequential analysis.
Solution Approach 2:
The patent introduces an intermediary processing step that automatically generates and presents parameter influence information. This intermediary layer translates the single estimation value into detailed parameter-specific insights without requiring the user to manually perform multiple inference operations, thus preserving both efficiency and information quality.
2Loss of information
If multiple data sets are prepared to understand parameter influence, then detailed parameter impact information is obtained, but user convenience is lowered and processing efficiency deteriorates
Solution Approach 1:
The system performs self-service by automatically generating parameter influence information through automated inference processing. Instead of requiring users to manually prepare multiple data sets and perform repeated analysis, the system autonomously conducts the necessary processing and presents the results, significantly improving user convenience while maintaining information completeness.
Solution Approach 2:
The patent performs preliminary action by automatically executing the multiple inference operations needed to analyze parameter influence before the user needs the information. This preliminary processing eliminates the need for users to understand or perform complex repeated analysis, making the system easier to use while providing comprehensive parameter impact data.
3Measurement precision
If the number of parameters input to the neural network increases, then more comprehensive performance estimation is achieved, but the difficulty of understanding parameter influence increases
Solution Approach 1:
The patent applies local quality by providing specific, localized information about each parameter's individual influence on the predicted performance. Instead of presenting a single complex multi-parameter estimation, the system breaks down the influence into discrete parameter-specific values, making it easier for users to understand the contribution of each parameter while maintaining comprehensive performance estimation accuracy.
Data Source
AI summary
A data processing apparatus according to one aspect is provided with an inference unit configured to predict at least one objective variable from a plurality of explanatory variables by using a trained model, and a display data generation unit configured to generate data for displaying an inference result by the inference unit. The inference unit is configured to set a first explanatory variable selected from the plurality of explanatory variables as a variation value and set second explanatory variables other than the first explanatory variable as fixed values. The inference unit predicts, by using the trained model, the at least one objective variable when the first explanatory variable is continuously varied within a predetermined variation range. The display data generation unit generates data indicating a variation of the at least one objective variable with respect to a variation of the first explanatory variable.