Streaming Decision Tree Updates for Anomalous Automation Outputs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control systems using decision trees for monitoring and diagnostics in industrial automation processes often output anomalous classifications, requiring retraining of the entire decision tree, which is inefficient and resource-intensive.
Innovation Solution
The system modifies a portion of the decision tree associated with anomalous outputs, rather than retraining the entire tree, by adjusting thresholds, splitting nodes, rearranging branches, or adding new nodes, allowing for incremental updates and improved robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the entire decision tree is retrained when anomalous outputs are detected, then the accuracy and reliability of the decision tree are improved, but the processing time and computational resources are significantly increased
Solution Approach 1:
The patent segments the decision tree into multiple modules or nodes that can be independently identified and modified. When an anomalous output is detected, only the specific module or node causing the anomaly is retrained rather than the entire decision tree, thereby reducing retraining time while maintaining accuracy.
Solution Approach 2:
The patent extracts and identifies the specific module or node within the decision tree that is producing anomalous outputs. By isolating and targeting only this problematic segment for retraining, the system avoids the computational overhead of retraining the entire decision tree while still improving reliability.
2Reliability
If the entire decision tree is retrained when anomalous outputs are detected, then the reliability of the decision tree is improved, but the computational resources and processing power are significantly increased
Solution Approach 1:
The patent segments the decision tree into multiple modules or nodes that can be independently identified and modified. When an anomalous output is detected, only the specific module or node causing the anomaly is retrained rather than the entire decision tree, thereby reducing computational resources while maintaining accuracy.
Solution Approach 2:
The patent applies partial action by retraining only the necessary portion of the decision tree (the specific module or node causing anomalies) rather than performing excessive action by retraining the entire decision tree. This partial retraining approach conserves computational resources while still achieving the desired reliability improvement.
3Adaptability or versatility
If the entire decision tree is retrained when anomalous outputs are detected, then the decision tree adapts to new conditions, but the operational efficiency and productivity are reduced
Solution Approach 1:
The patent segments the decision tree into multiple modules or nodes that can be independently identified and modified. When an anomalous output is detected, only the specific module or node causing the anomaly is retrained rather than the entire decision tree, thereby maintaining adaptability while minimizing disruption to operational efficiency.
Solution Approach 2:
The patent implements a dynamic approach where the decision tree structure allows for selective modification of specific modules or nodes based on detected anomalies. This dynamic adaptability enables the system to respond to changing conditions efficiently without requiring complete retraining, thus maintaining productivity.
Data Source
AI summary
A non-transitory computer-readable medium comprising computer-executable instructions that, when executed, are configured to cause a processor to perform operations that include receiving operational parameters for one or more automation devices, wherein the one or more automation devices are configured to implement control logic generated based on a decision tree. The operations also include receiving an output by the decision tree based on the operational parameters. Further, the operations include determining the output is an anomalous output based on a constraint associated with the decision tree. Further still, the operations include generating an updated decision tree based on the anomalous output. Even further, the operations include generating updated control logic for the one or more automation devices based on the updated decision tree. Even further, the operations include sending the updated control logic to the one or more automation devices.