Anomaly Score Drift Alarms for Multi-Class Input Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software management systems for anomaly detection in industrial environments face challenges in providing efficient alarms related to anomaly scores, requiring advanced coding knowledge and involving manual, time-consuming processes.
Innovation Solution
A computer-implemented method and system that receives input data from devices, applies anomaly detection models to generate output data, and provides an alarm if the difference between determined anomaly scores exceeds a threshold, simplifying the process through a user-friendly interface and automated detection of data distribution drift.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual processes and advanced coding knowledge are used for anomaly detection alarm provision, then the system can provide detailed control and customization, but the process becomes time-consuming and complex
Solution Approach 1:
The system performs automated anomaly detection and alarm provision without requiring manual intervention. The anomaly detection model automatically analyzes input data, determines anomaly scores, compares them against thresholds, and provisions alarms when anomalies are detected, enabling the system to serve itself rather than requiring human operators for each alarm provision task
Solution Approach 2:
The patent replaces manual mechanical processes with an automated computational system. Instead of human operators manually analyzing data and provisioning alarms, an anomaly detection model using machine learning automatically performs these tasks, substituting the mechanical human operation with an automated electronic system that processes data and generates alarms
2Adaptability or versatility
If manual processes and advanced coding knowledge are required for anomaly detection, then the system can provide customized alarm provisions, but the ease of operation decreases
Solution Approach 1:
The system automatically determines anomaly scores and provisions alarms based on pre-configured thresholds and anomaly detection models, eliminating the need for users to manually code or configure complex detection logic. The system serves itself by automatically adapting to data patterns and provisioning alarms according to established criteria
Solution Approach 2:
The anomaly detection model serves multiple functions: it analyzes input data, determines anomaly scores, compares scores against thresholds, and provisions alarms. This multi-functional approach allows the system to handle various anomaly detection scenarios with a single unified model, providing adaptability without requiring users to operate multiple specialized tools or code custom solutions
3Productivity
If automated anomaly detection is implemented, then the productivity and efficiency improve, but the system complexity increases
Solution Approach 1:
The patent extracts the complex anomaly detection logic into a separate, dedicated anomaly detection model that operates independently from the main system operations. This extracted model can be trained and updated separately, allowing the main system to benefit from automated detection without bearing the full complexity of model training and management, thereby improving productivity while managing system complexity
Data Source
AI summary
For improved provision of an alarm relating to anomaly scores assigned to input data, a method includes receiving input data relating to at least one device. The input data includes incoming data batches X relating to at least N separable classes. Respective anomaly scores are determined for the respective incoming data batch X relating to the at least N separable classes using N anomaly detection models. The anomaly detection models are applied to the input data to generate output data. A difference is determined, for the respective incoming data batch X, between the determined respective anomaly scores for the at least N separable classes and given respective anomaly scores of the N anomaly detection models. When the respective determined difference is greater than a difference threshold, an alarm relating to the determined difference is provided to a user, the respective device, and/or an IT system connected to the respective device.