Dynamic Weight Ensemble Model for Biometric Prediction Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ensemble models assign the same weight values to result values from multiple prediction models, leading to reduced uncertainty and reliability in final results, especially when noise or unlearned data is input, as they fail to account for the varying uncertainties of individual models.
Innovation Solution
A computer-implemented method that determines weight values for each model's result based on its uncertainty, using sensing information and uncertainty determination models to combine the weighted results and produce a final output, thereby addressing the uncertainty differences among models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the same weight value is assigned to result values from multiple prediction models, then the ensemble model is simple to implement, but the reliability of the final result decreases when noise or unlearned data is input
Solution Approach 1:
The patent changes the parameter of weight values from fixed (same for all models) to dynamic (different for each model based on uncertainty). Each model's weight is adjusted according to its predicted uncertainty, allowing the ensemble to adaptively respond to noise and unlearned data while maintaining implementation feasibility through automated uncertainty calculation.
Solution Approach 2:
The patent introduces dynamics into the weight assignment process by calculating uncertainty dynamically for each model based on its input data and predictions. This allows the ensemble model to automatically adjust the contribution of each prediction model according to the quality and reliability of its output, improving overall reliability without manual intervention.
2Reliability
If different weight values are assigned to result values from multiple prediction models based on uncertainty, then the reliability of the final result improves, but the complexity of the ensemble model increases
Solution Approach 1:
The patent implements self-service by enabling each prediction model to self-evaluate its own uncertainty based on its input data and predictions. The ensemble model automatically calculates appropriate weight values using these self-assessed uncertainties, eliminating the need for manual weight tuning or complex external evaluation mechanisms, thus improving reliability while controlling complexity.
3Measurement precision
If uncertainty calculation is performed for each model using input and output values, then the accuracy of weight determination improves, but the computational cost increases
Solution Approach 1:
The patent replaces complex mechanical or manual uncertainty assessment mechanisms with computational methods that calculate uncertainty directly from model inputs and outputs. This substitution enables accurate weight determination through automated mathematical calculations rather than manual evaluation, improving precision while managing computational resources efficiently through algorithmic optimization.
Data Source
AI summary
A computer-implemented method for controlling a device based on an ensemble model can include receiving sensing information associated with a user's biometric state; inputting first sensing information to a first model, determining a first uncertainty of the first model, and generating a first weight value for weighting a first result value; inputting second sensing information into a second model, determining a second uncertainty of the second model, and generating a second weight value for weighting a second result value; generating a final result value based on combining the first result value weighted by the first weight value and the second result value weighted by the second weight value; generating a predicted biometric state of the user based on the final result value; and executing an operation of the device based on the predicted biometric state.