Rule-Chain Data Processing for Interpretable Predictions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current neural network models lack the ability to provide a clear basis for their output results, limiting their application in fields requiring transparent prediction processes and reliable judgments, such as medical care, finance, and education.
Innovation Solution
A data processing method and apparatus that utilizes a prediction model with multiple rule chains, each having a corresponding prediction result and analysis basis, allowing for the determination of a target prediction result and its basis by matching attribute data with the analysis basis of a target rule chain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a neural network model is used for prediction, then prediction capability is improved, but interpretability and transparency of the prediction process deteriorate
Solution Approach 1:
The patent segments the prediction model into multiple independent rule chains, where each rule chain contains a specific prediction rule and its corresponding analysis basis. This segmentation allows the system to maintain the predictive power of complex models while providing transparent, interpretable reasoning paths for each prediction, thus resolving the contradiction between prediction accuracy and interpretability.
Solution Approach 2:
The patent introduces an intermediary component that connects the prediction model with the analysis basis. This intermediary captures and preserves the reasoning process between input data and prediction output, enabling the system to provide both accurate predictions and interpretable bases without compromising either aspect.
2Measurement precision
If a neural network model is used for prediction, then prediction capability is improved, but reliability in fields requiring transparent criteria deteriorates
Solution Approach 1:
By segmenting the model into rule chains with explicit prediction rules and analysis bases, the system enables reliable auditing and verification in transparent fields while maintaining prediction accuracy. Each rule chain can be independently validated, enhancing overall system reliability.
Solution Approach 2:
The patent makes the previously invisible reasoning process visible by introducing analysis bases that explain the prediction logic. This 'coloring' of the black-box model with interpretable reasoning paths enables reliable application in fields requiring transparency without sacrificing prediction capability.
3Loss of information
If analysis basis is added to prediction model, then interpretability is improved, but device complexity increases
Solution Approach 1:
The patent organizes the complex system into modular rule chains, where each chain is a self-contained unit with a specific prediction rule and analysis basis. This modular segmentation makes the overall complex system manageable and easier to maintain, reducing the practical complexity despite adding interpretability features.
Solution Approach 2:
The rule chain structure serves multiple functions simultaneously: it performs prediction, provides interpretation, enables auditing, and facilitates maintenance. This multi-functionality reduces the need for separate components, thereby limiting the increase in overall device complexity while achieving improved interpretability.
Data Source
AI summary
The present disclosure provides a data processing method, apparatus and electronic device. The data processing method includes: acquiring (S201) attribute data of a target object, where the target object includes one of an image, a text, a voice or a user; inputting (S202) the attribute data into a prediction model for analysis to obtain a target prediction result corresponding to the attribute data and a target analysis basis for obtaining the target prediction result, where the prediction model includes a plurality of rule chains, each of which has a corresponding prediction result and an analysis basis, and the target prediction result is determined according to a prediction result corresponding to a target rule chain, and the target analysis basis is determined according to an analysis basis corresponding to the target rule chain, and the attribute data meets the analysis basis corresponding to the target rule chain.