Tree Structure Models for Anomaly Detection Under Concept Drift

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing anomaly detection techniques struggle to adapt to changes in data trends over time, leading to inaccurate classifications due to concept drift or insufficient data, where normal states are misidentified as anomalous and vice versa.

Innovation Solution

Anomaly detection apparatus utilizing tree structure models that incorporate additional learning and forgetting learning processes to update models based on new data, enabling adaptation to changing data trends.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a tree structure model is used for anomaly detection, then anomaly detection can be performed, but the model cannot adapt to changing data trends over time

Engineering Contradiction:

Improveanomaly detection accuracyVSAvoidadaptability to data trend changes

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic model updating by introducing a forgetting learning process that periodically adjusts the tree structure model parameters. This allows the model to adapt to changing data trends while maintaining its anomaly detection capability, resolving the contradiction between reliability and adaptability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the model parameters by performing additional learning with new data and forgetting learning to update the tree structure model. This parameter updating mechanism enables the model to adapt to concept drift and changing data distributions, maintaining detection accuracy over time.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If the model is updated frequently with new data, then adaptability to data trends improves, but the complexity of the model increases

Engineering Contradiction:

Improveadaptability to data trendsVSAvoidmodel update complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts the updating mechanism from the main model by separating additional learning and forgetting learning processes. This allows the model to be updated in a modular manner, reducing the complexity of model maintenance while maintaining adaptability to data trends.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent implements feedback mechanisms through the forgetting learning process that uses evaluation target data to adjust model parameters. This feedback loop enables automatic adaptation to data trends without requiring complex manual intervention, balancing adaptability with operational simplicity.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20250278640A1Anomaly detection apparatus, system, method, and program

Publication Date: 2025.09.04 AISING LTD

AI summary

An anomaly detection apparatus includes: data acquisition processor circuitry acquiring evaluation target data; an anomaly detector performing anomaly detection about the evaluation target data based on an inference output generated based on input of the evaluation target data to one or more tree structure models; and an updated model generator generating one or more updated tree structure models by performing an additional learning process based on the evaluation target data and a forgetting learning process for the one or more tree structure models; wherein if a predetermined condition is satisfied, the one or more updated tree structure models are used instead of the one or more tree structure models in the anomaly detector and the updated model generator.