Incident Summarization With Symptom-Resource Pairing Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing frameworks for generating incident reports from multi-modal data in computer systems are cumbersome, time-consuming, manually driven, and inefficient, making it difficult to derive meaningful insights from heterogeneous data sources.
Innovation Solution
A computer-implemented method using natural language processing and generative machine learning models to automatically generate abstractive summaries of IT incidents, incorporating symptom-resource pairings and user-specific preferences, while addressing AI hallucinations through feedback mechanisms and AIOps-specific metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual frameworks are used to generate incident reports from multi-modal data, then detailed analysis can be performed, but the process becomes cumbersome and time-consuming
Solution Approach 1:
The patent replaces manual mechanical processes with an automated machine learning system. Specifically, it uses a trained machine learning model that automatically ingests multi-modal incident data (logs, metrics, traces), performs analysis, and generates natural language reports without human intervention, thereby eliminating the time-consuming manual framework while maintaining analytical depth
Solution Approach 2:
The system enables self-service incident analysis by automatically processing incident data through the trained machine learning model. The model autonomously performs data ingestion, analysis, report generation, and even self-evaluation through feedback mechanisms, eliminating the need for manual operation while delivering comprehensive incident insights
2Adaptability or versatility
If manual processes are used for incident summarization, then customization to user preferences is possible, but manual effort and complexity increase
Solution Approach 1:
The patent implements dynamic customization where the system adapts to user preferences automatically. The machine learning model generates reports that can be customized to different user roles and preferences, and the system dynamically adjusts based on feedback received during and after report generation, eliminating manual configuration complexity while maintaining high adaptability
Solution Approach 2:
The system performs preliminary configuration by pre-training the machine learning model on diverse incident data and pre-configuring report templates that accommodate various user preferences. This preliminary preparation enables the system to automatically adapt to different users without requiring manual setup, reducing process complexity while maintaining versatility
3Productivity
If automated summarization is implemented, then efficiency improves, but accuracy and handling of AI hallucinations become challenges
Solution Approach 1:
The patent implements a feedback mechanism where the generated natural language reports are evaluated against the original incident data and ground truth. The machine learning model receives feedback on its performance, including identification of hallucinations or inaccuracies, and uses this feedback to refine and improve subsequent report generation, thereby maintaining high accuracy while preserving automation efficiency
Solution Approach 2:
The system performs preliminary validation by cross-checking generated summary content against the original multi-modal incident data before finalizing the report. This preliminary anti-action prevents hallucinations and inaccuracies from being propagated, ensuring high accuracy while maintaining automated efficiency
4Loss of information
If comprehensive multi-modal data is analyzed, then insight quality improves, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary data preparation by pre-processing and structuring multi-modal incident data (logs, metrics, traces) before the actual analysis. The machine learning model is pre-trained on diverse incident data, enabling it to efficiently process comprehensive data during incident analysis without excessive processing time, thereby maintaining insight quality while reducing computational overhead
Data Source
AI summary
A computer-implemented method includes parsing, by a processor set, text comprised by an alert corresponding to an information technology (IT) abnormality incident, resulting in alert data. The processor set uses a generative machine learning (ML) model to generate a natural language summary of the incident. The natural language summary includes a symptom-resource pairing corresponding to the alert and is based on the alert data and on a topology of keywords comprised by the alert. In one or more embodiments, the computer-implemented method further comprises employing, by the processor set, graph connectivity distances between elements of the topology to verify the symptom-resource pairing.