Behavioral Alert Directives for Consistent Event Publishing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Behavioral recognition systems struggle to provide consistent alerts for specific events without disrupting their unsupervised learning process, as they rely solely on unsupervised learning and lack predefined rules for alerting on certain behaviors.
Innovation Solution
Implementing alert directives that allow users to define criteria for specific events to always or never result in an alert, bypassing the normal alert publication process of the machine learning engine, thus maintaining the unsupervised learning integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a behavioral recognition system relies solely on unsupervised learning to determine alerts, then the system maintains learning integrity and adaptability, but the system cannot provide consistent alerts for specific predefined events
Solution Approach 1:
The system segments the alert determination process into two independent components: (1) unsupervised learning for anomaly detection that maintains adaptability, and (2) predefined alert directives for consistent alerting on specific events. This segmentation allows each component to fulfill its specialized function without compromising the other.
Solution Approach 2:
Alert directives serve as an intermediary layer between the unsupervised learning engine and the alert publication process. These directives act as predefined rules that override the learning-based anomaly determination for specific events, ensuring consistent alerting while preserving the integrity of the unsupervised learning process.
2Reliability
If predefined rules are added to ensure consistent alerting, then alert reliability improves, but the system complexity increases
Solution Approach 1:
Alert directives are configured in advance as predefined rules before the system operates. This preliminary configuration allows the system to have consistent alerting behavior for specific events without adding complex real-time processing logic, as the decision rules are already established beforehand.
Solution Approach 2:
The system uses a simplified copy of the anomaly detection logic embedded within alert directives. Rather than creating a completely separate rule-based system, the alert directives replicate the essential alert determination criteria in a simplified, predefined format that is easier to manage and less complex than full behavioral recognition rules.
3Ease of operation
If the system uses unsupervised learning alone, then the system remains simple to operate, but the system cannot alert on specific events that should always or never result in alerts
Solution Approach 1:
Instead of requiring complete reconfiguration of the unsupervised learning system to add alert control, the invention applies a partial solution through alert directives. These directives provide just enough additional control for specific events where consistent alerting is needed, while leaving the rest of the system operating in its simple unsupervised learning mode.
Data Source
AI summary
Alert directives and focused alert directives allow a user to provide feedback to a behavioral recognition system to always or never publish an alert for certain events. Such an approach bypasses the normal publication methods of the behavioral recognition system yet does not obstruct the system's learning procedures.