Anomaly Detection via Access Data Probability Differences

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current intrusion detection systems face challenges in effectively identifying and responding to sophisticated attacks on computer systems, as they often rely on static patterns and may not adapt well to changing access data probabilities.

Innovation Solution

A method that involves logging access data, performing anomaly analysis by identifying combinations of access parameters with significant probability differences within specific time intervals, and monitoring these parameters to detect new attacks, utilizing predefined probability differences to trigger countermeasures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If static patterns are used for attack detection, then the detection system is simple to implement, but it cannot adapt to changing access data probabilities and sophisticated attacks

Engineering Contradiction:
Improveadaptability to changing attack patternsVSAvoidcomplexity of detection system
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic anomaly detection by continuously calculating probability differences of access parameter combinations compared to learned normal behavior patterns. Instead of using fixed static thresholds, the system dynamically adapts to changing access patterns by comparing current access probabilities against historically learned probabilities, enabling the system to adapt to sophisticated and evolving attacks while maintaining reasonable computational complexity through focused probability analysis.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the detection parameter from static pattern matching to dynamic probability difference analysis. By monitoring how probability distributions of access parameter combinations change over time and comparing them against learned baselines, the system can detect deviations indicating attacks without requiring complex adaptive algorithms, thus resolving the contradiction between adaptability and complexity.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If comprehensive access data analysis is performed to detect sophisticated attacks, then detection accuracy improves, but analysis time and computational resources increase

Engineering Contradiction:
Improveattack detection accuracyVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts and focuses analysis on specific probability difference metrics rather than performing comprehensive analysis of all access parameters. By identifying and monitoring key probability differences between normal and anomalous access patterns, the system achieves high detection accuracy while significantly reducing computational overhead and analysis time compared to analyzing all possible access data combinations.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system performs partial analysis by focusing on probability differences of specific access parameter combinations rather than analyzing all access data comprehensively. This selective approach to probability monitoring enables timely detection of sophisticated attacks by concentrating computational resources on the most indicative probability deviations, thus balancing accuracy with analysis speed.

Inventive Principle:
Principle #16Partial or excessive action

3Loss of time

If probability difference thresholds are set to be sensitive for early detection, then detection timing improves, but false positive rate increases

Engineering Contradiction:
Improvedetection timingVSAvoidfalse positive rate
Core Design Contradiction:
Loss of timeVSReliability

Solution Approach 1:

The system employs feedback mechanisms by continuously comparing current access probability differences against learned baseline probabilities and adjusting detection decisions based on cumulative evidence. This feedback loop allows the system to maintain sensitive detection thresholds for early attack detection while using historical probability patterns to filter out anomalies that deviate from typical attack signatures, thereby reducing false positives through adaptive threshold adjustment based on observed probability distributions.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4345665A1Detecting an attack on a computer system to be protected
Publication Date: 2024.04.03 BUNDESDRUCKEREI GMBH
  • EP4345665A1 patent drawingFigure 1
  • EP4345665A1 patent drawingFigure 2
  • EP4345665A1 patent drawingFigure 3

AI summary

The invention relates to a method for detecting an attack (110) on a computer system (100, 130, 160, 198) to be protected using logged access data (122, 152, 182) of accesses to the computer system (100, 130, 160, 198) to be protected. The procedure includes logging access data (122, 152, 182), performing an anomaly analysis for the occurrence of an attack on the computer system to be protected (100, 130, 160, 198), identifying those access parameters (112) whose values ​​are encompassed by one or more specific anomalies, and monitoring the values ​​logged in the access data records for the identified access parameters (112), the monitoring including the detection of a renewed attack (110).