Hybrid Explainable AI System Using Shallow Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing artificial intelligence systems, particularly deep neural networks, struggle to provide explanations for their decision-making processes, leading to inefficiencies and the need for human intervention to recreate the processes.
Innovation Solution
A hybrid explainable artificial intelligence system is developed, combining a shallow learning model based on maximum entropy with a deep learning model. This system inputs data into both models, uses common outputs to formulate explanations for the deep learning model, and provides visual representations to enhance understanding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a deep neural network is used to improve prediction accuracy, then the model's predictive capability is enhanced, but the system becomes incapable of generating explanations for its decision-making process
Solution Approach 1:
The system segments the explanation generation task from the deep neural network prediction task. A separate shallow learning model is trained to predict the same target variable, and its interpretable features are used to explain the deep network's predictions, allowing both accuracy and explainability to coexist
Solution Approach 2:
The shallow learning model acts as an intermediary between the deep neural network and the user. It translates the complex, opaque predictions of the deep network into human-understandable explanations based on its own interpretable feature importance, without sacrificing the deep network's predictive accuracy
2Loss of information
If human intervention is used to recreate and explain the neural network process, then explanation accuracy is improved, but resource consumption increases significantly
Solution Approach 1:
The system uses the shallow learning model to automatically generate explanations for the deep neural network's predictions without requiring human intervention. The model self-explains by leveraging its own interpretable feature importance, significantly reducing resource consumption while maintaining explanation quality
Solution Approach 2:
Instead of having humans manually recreate the neural network process, the system creates a simplified copy (shallow learning model) that mimics the deep network's predictions but with interpretable features. This copy serves as both a predictor and an explainer, eliminating the need for resource-intensive human analysis
Data Source
AI summary
A hybrid explainable artificial intelligence system may include a shallow learning model and a deep learning model. The shallow learning model may be a machine learning system. The deep learning model may be a neural network. The system may input a data set into both the shallow learning model and the deep learning model. Both the shallow learning model and the deep learning model may produce an output. When there is a common output between the shallow learning model and the deep learning model, the process performed by the shallow learning model may be used to formulate an explanation of the process performed by the deep learning model. The explanation of the process performed by the deep learning model may be used to raise the sensitivity of one or more components of the data set. Such components may include a word or phrase within a transcript.