Logic-Based Neural Networks With Rules Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional neural networks lack accuracy when processing inputs significantly different from training data and fail to utilize contextual information, leading to incorrect outputs without providing explanations for their decisions.
Innovation Solution
An AI model that incorporates a rules engine to process outputs from neural networks using both feature data and external information, generating more accurate results and including contextual explanations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a neural network is trained using conventional methods, then it can process inputs similar to training data, but it fails to provide accurate output when given inputs significantly different from training data
Solution Approach 1:
The patent introduces a rules engine as an intermediary component between the neural network and the final output. The rules engine receives the neural network's output along with contextual information from external data sources, applies logical rules to evaluate and adjust the output, and generates a final result. This mediator resolves the contradiction by allowing the system to maintain high accuracy on training-like inputs while adapting to handle novel inputs through rule-based reasoning and contextual awareness.
2Productivity
If a neural network processes input in isolation, then it generates output quickly, but it cannot utilize contextual or external data to improve accuracy
Solution Approach 1:
The patent segments the processing system into distinct functional components: the neural network handles rapid pattern recognition and generates preliminary output, while the rules engine handles contextual analysis and final decision-making. This segmentation allows each component to specialize - the neural network maintains high processing speed for initial analysis, while the rules engine provides accuracy through structured reasoning over multiple data sources including external contextual information.
3Device complexity
If a neural network generates output without contextual information, then it operates simply, but the output lacks explanations that can guide decisions or analysis
Solution Approach 1:
The rules engine serves as an intermediary that not only processes the neural network's output but also generates explanatory information. It evaluates rules, accesses external data sources for contextual information, and produces outputs that include both the final result and the reasoning behind it. This maintains the relative simplicity of the neural network while adding the missing contextual explanations through the rules engine's transparent rule-based reasoning process.
Data Source
AI summary
Various embodiments set forth systems and techniques for augmenting neural networks. The techniques include causing one or more neural networks to generate first output based on a first input; identifying one or more rules associated with the first input; processing the first output based on the one or more rules to generate a second output; and transmitting the second output, instead of the first output, as a result of processing the first input.