GRU Neural Network Root Cause Prediction for Computer Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As computer systems become increasingly complex, existing monitoring tools struggle to accurately identify the root cause of errors due to multiple potential causes, leading to misidentification or failure in identifying the root cause.
Innovation Solution
A computer-implemented method using a gated recurrent unit (GRU) neural network that retrieves performance data from a target computer system to predict the root cause of an error alert, with the weights of the GRU neural network adjusted based on user feedback to improve prediction accuracy over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If static rules-based analysis is used for monitoring, then the system is simple to implement, but the accuracy of root cause identification deteriorates as computer systems become more complex
Solution Approach 1:
The patent transforms the monitoring system from static rules-based analysis to dynamic neural network-based analysis. The neural network continuously learns from new data patterns, allowing the system to adapt its analysis parameters automatically. This resolves the contradiction by enabling high accuracy without requiring complex manual rule configuration, as the system self-adjusts through training on system performance data.
Solution Approach 2:
The patent implements feedback mechanisms where the neural network receives feedback about its predictions and systematically adjusts its internal parameters accordingly. This feedback loop allows the system to improve its root cause identification accuracy over time while maintaining ease of use, as the feedback is processed automatically without requiring manual intervention or complex configuration.
2Measurement precision
If neural network-based analysis is used to improve root cause identification accuracy, then the accuracy improves, but the device complexity increases
Solution Approach 1:
The patent implements self-service through automated neural network training and parameter adjustment. The system automatically retrieves performance data, trains the neural network on new patterns, and adjusts its parameters without requiring manual intervention. This self-service capability achieves high accuracy while managing complexity automatically, as the system handles its own optimization rather than requiring complex external management.
Solution Approach 2:
The patent extracts the complexity management into a separate neural network component that handles pattern recognition independently from the core monitoring functions. By separating the intelligent analysis layer from the data collection layer, the system achieves high accuracy through the neural network while keeping the overall architecture manageable through modular design.
3Measurement precision
If the system analyzes multiple potential causes simultaneously, then the completeness of root cause analysis improves, but the time required for analysis increases
Solution Approach 1:
The patent replaces mechanical sequential analysis with neural network-based parallel pattern recognition. The neural network processes multiple potential causes simultaneously by analyzing different features and patterns in parallel, achieving both completeness and speed. This substitution of mechanical sequential processing with neural network parallel processing resolves the time-completeness contradiction.
Solution Approach 2:
The patent adds the dimension of temporal dynamics to the analysis by continuously updating the neural network with new performance data over time. This temporal dimension allows the system to learn from historical patterns and make faster, more accurate root cause identifications without requiring exhaustive sequential analysis of all possible causes, as the neural network has already learned from past data.
Data Source
AI summary
Aspects of the invention include detecting an error alert from a target computer system. In response to detecting the error alert, performance data is then retrieved from the target computer system. A gated recurrent unit (GRU) neural network is used to generate a prediction of a root cause of the error alert based on the performance data. The weights of a reset gate of the GRU neural network are adjusted based on received feedback of the prediction.