Coordinate Vector Security Violation Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current security systems struggle to effectively detect and differentiate between malicious activities, including malware and non-malware threats, often missing sophisticated security violations such as unauthorized access or data exfiltration, due to reliance on signature-based detection methods that are inadequate against polymorphic malware and zero-day exploits.

Innovation Solution

A method involving the classification of events on monitored computing devices using coordinate vectors and machine learning techniques to identify potential security violations, allowing for the correlation of data from multiple events and the presentation of these events in a user-friendly interface to distinguish between clean and malicious activities, thereby reducing false positives and negatives.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If signature-based detection methods are used, then known malware can be detected, but sophisticated security violations such as polymorphic malware and zero-day exploits are missed

Engineering Contradiction:
Improvedetection accuracyVSAvoidability to detect new threats
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent segments malicious activities into distinct categories (malware, non-malware, targeted attacks) and represents each category using specific coordinate vectors with targeted features. This segmentation allows the system to detect different types of threats using appropriate features for each category, improving both reliability for known threats and adaptability for new threats.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms security event data into coordinate vectors that represent activities in a multi-dimensional space. By mapping events to coordinates based on multiple features simultaneously, the system can detect threats from multiple dimensions rather than relying on single signature matching, enabling detection of both known and unknown threats.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If multiple features are analyzed to improve detection accuracy, then false positives may increase, but reducing features improves speed

Engineering Contradiction:
Improvedetection precisionVSAvoidprocessing speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent applies different feature sets (coordinate vectors) to different types of security events based on their specific characteristics. Each event type is evaluated using locally optimized features relevant to that category, improving detection precision without unnecessarily analyzing all possible features for every event, thus maintaining processing speed.

Inventive Principle:
Principle #3Local quality

3Reliability

If comprehensive event monitoring is implemented, then more security violations can be detected, but the volume of alerts for analysts increases

Engineering Contradiction:
Improvesecurity violation detectionVSAvoidalert volume
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts and highlights only the most significant security violations by representing events as coordinate vectors and identifying those that deviate from normal patterns. This extraction approach allows comprehensive monitoring while presenting only the most critical alerts to analysts, reducing alert volume without compromising detection reliability.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3716111B1Computer-security violation detection using coordinate vectors
Publication Date: 2023.02.22 CROWDSTRIKE
  • EP3716111B1 patent drawingFigure 1
  • EP3716111B1 patent drawingFigure 2
  • EP3716111B1 patent drawingFigure 3

AI summary

Event vectors can be determined for respective events based on respective command-line records and a trained representation mapping. Respective coordinate vectors can be determined, each having fewer elements than the respective event vector. Respective representations of at least some of the events can be presented via an electronic display at the respective coordinate vectors. A selection of a first representation can be received via a user interface. The events can be clustered based on the event vectors. A first cluster can be selected based on the selection. An indication of a tag can be received via the user interface. Each event of the first cluster can be associated with the tag. Some examples include transmitting a security command to cause a monitored computing device associated with an event in the first cluster to perform a mitigation action.