AI Model Quality Improvement Through Change-Risk Incident Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately assessing change risk and impact, leading to increased incidents and outages, particularly in complex IT environments, due to subjective and static methods for evaluating software and hardware modifications.
Innovation Solution
A machine-learning based model is used to generate queries for change descriptions, risks, and implementation plans, receiving owner responses to refine the model's association between change features and incident features, providing intelligent alerts and predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static surveys or guessing are used to assess change risk and impact, then the assessment process is simple and quick, but the accuracy and reliability of risk assessment deteriorates
Solution Approach 1:
The patent replaces manual static surveys and guessing with an automated machine learning model that analyzes historical incident data, change data, and system data to predict change-caused incidents. The ML model processes multiple data sources simultaneously to generate accurate risk assessments without manual intervention, resolving the contradiction between simple assessment processes and accurate risk prediction.
Solution Approach 2:
The patent introduces an intermediary machine learning model that acts as a bridge between raw data (historical incidents, change requests, system configurations) and risk assessment outcomes. This intermediary processes and synthesizes multiple data sources to provide accurate, data-driven risk predictions, eliminating the need for subjective guessing while maintaining assessment efficiency.
2Ease of manufacture
If manual rules-based approaches are used to determine allowed changes, then implementation is straightforward, but the system cannot adapt to complex modern IT architectures
Solution Approach 1:
The patent implements a dynamic machine learning model that continuously learns from historical incident data and adapts to changing system configurations and patterns. Unlike static rules-based approaches, the ML model evolves with the system, automatically adjusting its predictions based on new data while maintaining ease of deployment through automated training and update mechanisms.
Solution Approach 2:
The patent creates a universal machine learning model that can handle diverse change types (software, hardware, configuration) and assess risks across modern complex IT architectures including cloud environments, microservices, and distributed systems. The single ML-based solution replaces multiple specialized rules, providing both simplicity and adaptability simultaneously.
3Measurement precision
If comprehensive data collection and ML model training are implemented, then prediction accuracy improves, but computational resources and time requirements increase
Solution Approach 1:
The patent performs preliminary data collection and ML model training during off-peak hours or in advance, preparing the model with historical incident data before it is needed for production risk assessment. This preliminary action allows the model to be pre-trained and optimized, reducing real-time computational requirements when actual change risk assessments are performed.
Solution Approach 2:
The patent implements self-service mechanisms where the ML model automatically collects relevant data, trains itself on new incident patterns, and updates its predictions without requiring extensive manual computational resources. The system autonomously manages its own learning and adaptation, reducing the energy and computational overhead required for maintaining high prediction accuracy.
Data Source
AI summary
A method for improving the quality of a machine-learning based model includes generating a first query requesting a description of a change proposed to a system and an intended outcome of the change proposed; receiving a first response; generating a second query providing a risk of an incident associated with the change proposed and requesting justification of the change proposed in view of the risk; receiving a second response; generating a third query requesting an implementation plan for the change proposed; receiving a third response; generating an alert to an incident owner providing the description, intended outcome, risk, justification, and implementation plan of the change proposed; receiving a risk confirmation or rejection from the incident owner confirming or rejecting a relationship between the change proposed and the risk; and updating the machine-learning based model to learn an association between extracted features of the change and extracted features of the incident.