Determination device, determination method, and determination program
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques face challenges in accurately determining operational information for newly developed apparatuses without a track record, as they rely on neural network models constructed for apparatuses with a track record, which may not directly apply due to differences in components and operation principles.
Innovation Solution
A determination device that includes a model information storage unit, an information acquisition unit, a determination unit, and a correction unit, which uses a neural network model constructed for a similar apparatus with a track record, and corrects input or output data using coefficients to ensure accurate operation of the new apparatus.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If a neural network model constructed for an apparatus with a track record is used to determine operational information for a newly developed apparatus, then the determination can be performed without collecting extensive data from the new apparatus, but the accuracy of determination deteriorates due to differences in components and operation characteristics between the two apparatus types
Solution Approach 1:
The patent transforms the neural network model by changing its parameters (input/output items and coefficients) to match the characteristics of the newly developed apparatus. This allows the model to be adapted from a proven apparatus to a new apparatus type, maintaining determination accuracy while avoiding extensive data collection from the new apparatus.
Solution Approach 2:
The patent creates a transformed version of the neural network model by copying the structure from an apparatus with a track record and then adapting it through parameter changes. This copied and transformed model can be directly applied to the new apparatus, eliminating the need to build a new model from scratch or collect extensive training data.
2Ease of manufacture
If a neural network model is transformed by changing input/output items and coefficients to match a newly developed apparatus, then the model can be directly applied without extensive data collection, but the complexity of model transformation increases
Solution Approach 1:
The patent systematically changes the parameters of the neural network model (input items, output items, and coefficients) to match the new apparatus characteristics. This structured parameter transformation approach makes the model adaptation process manageable and directly applicable, reducing overall system complexity despite the transformation steps involved.
3Measurement precision
If coefficients are used to correct input or output data from a transformed neural network model, then determination accuracy for the new apparatus is improved, but the complexity of data correction processes increases
Solution Approach 1:
The patent incorporates coefficient-based correction as an integral part of the neural network model transformation process. By systematically adjusting coefficients to match the new apparatus characteristics, the correction becomes a standardized parameter adjustment rather than a complex post-processing step, improving accuracy while maintaining manageable complexity.
Data Source
AI summary
A determination device includes: a model information storage unit for storing a neural network model constructed with respect to operation of a second apparatus which operates according to a same operation principle as a first apparatus to be controlled; an information acquisition unit for acquiring: first item regarding an apparatus which operates according to the operation principle; a first coefficient for calculating second item from the first item on the first apparatus, and a second coefficient for calculating second item from the first item on the second apparatus; a determination unit for determining the second item regarding the first apparatus, by using the neural network model, on the basis of the first item; and a correction unit for multiplying the first item or the second item, which is to be used by the determination unit, by a ratio of the first to second coefficients, thereby correcting corresponding data.