Visual Anomaly Detection Interface for APM Systems

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing anomaly detection techniques in Application Performance Management (APM) rely on numerical methods that lack accuracy in capturing human judgment, making them inefficient in identifying anomalies, especially in multivariate cases, as they require complex and tedious processes to incorporate human judgment.

Innovation Solution

A method that visually generates and modifies an anomaly detection image using a coordinate system with distinguishable areas, allowing users to define and associate actions with these areas, and generates formulas representing the areas, enabling intuitive detection and classification of anomalies through a graphical interface.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If numerical methods are used for anomaly detection, then the process can be automated, but the accuracy in capturing human judgment deteriorates

Engineering Contradiction:

Improveautomation of anomaly detectionVSAvoidaccuracy in capturing human judgment

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The patent introduces a visual coordinate system as an intermediary between human judgment and automated detection. Users interact with the visual interface to define distinguishable areas, which are then converted into mathematical formulas. This intermediary allows human expertise to be captured visually while maintaining automation through formula-based detection rules.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If complex numerical methods are used to incorporate human judgment, then detection accuracy may improve, but the process complexity increases

Engineering Contradiction:

Improvedetection accuracyVSAvoidprocess complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces complex numerical methods with a visual-coordinate-based system. Instead of using complicated algorithms to incorporate human judgment, users simply define areas on a visual coordinate system, which are then automatically converted into detection formulas. This substitution dramatically reduces process complexity while maintaining detection accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If visual methods are used for anomaly detection, then ease of operation improves, but automation extent may deteriorate

Engineering Contradiction:

Improveusability of detection systemVSAvoidautomation level

Core Design Contradiction:

Ease of operationVSExtent of automation

Solution Approach 1:

The visual coordinate system serves multiple functions: it provides an intuitive interface for user interaction, automatically converts user definitions into mathematical formulas, and generates executable detection rules. This multi-functionality allows the system to maintain high automation levels while preserving ease of operation through visual interaction.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10325386B2Visual generation of an anomaly detection image

Publication Date: 2019.06.18 CA TECH INC

AI summary

A method includes formatting for display, on a visual screen, an image comprising a coordinate system and a plurality of data points within the coordinate system, wherein the plurality of data points define historical performance data for a computer system. The method further includes receiving a user input defining a distinguishable area within the coordinate system and an action associated with the distinguishable area. The method additionally includes generating a formula representing the distinguishable area.