Score-Based Prediction Certainty Estimation With Adaptive Data Expansion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning systems produce inaccurate results when input data is thin (not large enough or not variant enough), leading to skewed predictions that do not match actual outcomes.
Innovation Solution
A machine learning system that expands input data by applying filters and gathering additional data until a confidence score exceeds a predetermined threshold, using a predetermined defocus scheme to increase data dimensionality and maintain prediction certainty.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If input data is filtered to be relevant for a particular prediction, then prediction accuracy is improved, but the quantity of input data decreases leading to thin data
Solution Approach 1:
The patent applies defocus schemes that expand data dimensionality by adding new features or attributes to the input data. This allows the system to maintain prediction accuracy while working with smaller datasets by creating more informative representations of the available data, effectively transforming the data structure rather than increasing raw quantity.
Solution Approach 2:
The system performs preliminary data expansion and defocusing operations before the actual prediction process. By pre-processing the data to add dimensional information and expand the feature space in advance, the system ensures that sufficient information is available for accurate predictions even when the original dataset is small.
2Quantity of substance
If input data is expanded by gathering additional data, then the quantity of input data increases, but the certainty of prediction decreases
Solution Approach 1:
The system changes parameters of the input data by applying defocus schemes that modify data characteristics, such as aggregating at different levels of granularity or transforming individual data points into grouped statistics. This allows expansion of data quantity while maintaining prediction certainty by changing how the data is structured and represented rather than simply adding more raw data.
Solution Approach 2:
The patent creates transformed copies of the original data through defocus operations, where data is replicated in different forms or aggregations. These copies provide additional information while maintaining the reliability of the original data, effectively increasing quantity without sacrificing certainty.
3Adaptability or versatility
If the dimension of input data is increased through defocus scheme, then data variance is improved, but the complexity of data processing increases
Solution Approach 1:
The patent segments the data processing into distinct stages: original data collection, defocus transformation, and prediction execution. By dividing the complex process into manageable segments with clear interfaces, the system can increase data variance through defocus schemes while keeping the overall complexity manageable through structured organization of processing steps.
Data Source
AI summary
The present disclosure describes a patent management system and method for remediating insufficiency of input data for a machine learning system. A prediction to be performed is received from a user input. Relevant input data is determined to perform the prediction. The relevant input data is determined by applying filters based on the prediction to be performed. Prediction is performed by generating a plurality of predicted vectors. A confidence score for the generated plurality of predicted vectors is determined. If the confidence score is less than a predetermined threshold, the prediction is unreliable. The input data is expanded by gathering additional input data. The input data is expanded with the additional input data until the confidence score exceeds the predetermined threshold. A predicted output is generated with the expanded input data. The prediction output and the confidence score are provided for rendering.