Dual Neural Network Predictor for Non-Explicit Parameter Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction methods for variables used in control systems, such as air traffic control, lack accuracy due to reliance on conventional statistical models and historical data, which fail to effectively account for non-explicit parameters.
Innovation Solution
A device comprising two neural network-based predictors that iteratively apply a learning function with forward propagation and backpropagation blocks to estimate non-explicit parameters and variable values, improving prediction accuracy by jointly training the networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional statistical models and historical data are used for prediction, then the prediction process is simple and fast, but the prediction accuracy is insufficient
Solution Approach 1:
The prediction system is segmented into two separate neural network predictors: a first predictor that processes initial parameters and a second predictor that processes the non-explicit parameter. This segmentation allows each predictor to specialize in specific parameter types, improving overall prediction accuracy while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The non-explicit parameter acts as an intermediary element that bridges the gap between observable parameters and the target variable. By explicitly modeling this intermediary parameter through a dedicated neural network predictor, the system can capture complex relationships that conventional statistical models miss, thereby improving prediction accuracy without overwhelming complexity.
2Reliability
If conventional statistical models are used, then the system is easier to implement, but it fails to account for non-explicit parameters effectively
Solution Approach 1:
The patent replaces conventional statistical models with neural network-based predictors that can automatically learn and adapt to complex parameter relationships. The first neural network predictor handles explicit parameters while the second handles non-explicit parameters, substituting traditional mechanical statistical methods with flexible neural network mechanisms that achieve higher reliability.
Solution Approach 2:
The system changes the approach to parameter handling by introducing a dedicated non-explicit parameter that is explicitly modeled by the second neural network predictor. This parameter transformation allows the system to reliably capture and account for previously unmodeled factors, improving prediction reliability while maintaining implementation feasibility through standardized neural network architectures.
3Measurement precision
If multiple predictors are used to handle different parameters, then prediction accuracy improves, but the learning and training process becomes more complex
Solution Approach 1:
The training process is segmented into independent training phases for each neural network predictor. The first predictor is trained on explicit parameters while the second predictor is trained on non-explicit parameters. This segmentation allows each predictor to be optimized independently, improving overall accuracy while simplifying the training process through modular, independent training procedures.
Solution Approach 2:
The system performs preliminary training of each neural network predictor separately before integrating their outputs for final prediction. This preliminary action allows each predictor to be thoroughly trained and optimized for its specific parameter type, improving accuracy while reducing the complexity of the overall training process by breaking it down into manageable preliminary steps.
Data Source
AI summary
The embodiments of the invention provide a device for predicting the value of a variable intended to be used by a computer-implemented control system, the variable depending on multiple parameters, the parameters comprising a non-explicit parameter. Advantageously, the prediction device comprises a first neural network-based predictor configured so as to compute an estimate of the non-explicit parameter and a second neural network-based predictor configured so as to compute an estimate of the value of the variable from the estimate of the non-explicit parameter, the two predictors receiving an input dataset, each neural network being associated with a set of weights. The prediction device is configured so as to apply a plurality of iterations of a single learning function to the two predictors, the learning function comprising: a forward propagation block for computing, on the basis of the input data of the two predictors, the gradient of a minimization function for minimizing a cost function of the first predictor; and a backpropagation block for updating the weights of the neural networks of the two predictors by backpropagating the gradients computed by the forward propagation block. The prediction device estimates the value of the variable to be predicted at a future time, after the iterations of the learning function, by applying input data to the neural networks of the two predictors and using the weights updated by the learning function.