ML Severity Assessment for Heterogeneous Network Disruptions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Communication networks in non-homogenous environments face challenges in detecting and responding to disruptions due to varying architectures, equipment, and terminology used by different users and systems, making it difficult to determine the severity of disruptions and coordinate responses effectively.
Innovation Solution
The use of multiple machine learning models, including a natural language processing model and a severity level determination model, to monitor network traffic, analyze user-generated data, and provide normalized severity level recommendations by comparing current and historic data across communication networks, enabling consistent disruption detection and response across diverse environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional disruption detection methods are used in non-homogenous communication networks, then the system can operate with simple monitoring, but the accuracy of disruption severity determination deteriorates due to varying architectures, equipment, and terminologies across different networks
Solution Approach 1:
The patent introduces machine learning models as intermediary components that translate and normalize disruption data from diverse network architectures, equipment types, and terminologies into a unified severity assessment framework. These models act as mediators between heterogeneous network elements and the central monitoring system, enabling accurate severity determination without requiring direct complex integration of all network variations.
Solution Approach 2:
The system transforms raw network disruption data into normalized severity parameters through machine learning processing. By changing the parameter representation from raw, heterogeneous network metrics to standardized severity levels, the system achieves accurate comparison and assessment across different network types while maintaining manageable system complexity.
2Measurement precision
If multiple machine learning models are deployed to analyze user-generated data and determine severity levels, then the disruption detection accuracy improves, but the computational resources and processing time increase
Solution Approach 1:
The system performs preliminary processing of user-generated data through machine learning models to extract and normalize relevant features before severity determination. By pre-processing data to identify key disruption indicators and normalize terminologies in advance, the system reduces the computational burden during real-time severity assessment while maintaining high detection accuracy.
Solution Approach 2:
The patent divides the disruption analysis process into multiple specialized machine learning models, each handling specific aspects such as data normalization, pattern recognition, and severity classification. This segmentation allows parallel processing of different data dimensions, improving overall detection accuracy while enabling efficient resource utilization through specialized computation for each sub-task.
3Reliability
If the system monitors and archives data according to predetermined time intervals across multiple communication networks, then the historical analysis capability improves, but the data storage requirements and processing overhead increase
Solution Approach 1:
The system extracts and archives only the most relevant disruption indicators and normalized severity metrics from raw network data according to predetermined time intervals. By selectively extracting key features rather than storing complete raw datasets, the system maintains reliable historical analysis capability while significantly reducing data storage requirements and processing overhead.
Solution Approach 2:
The system performs preliminary aggregation and normalization of network data before archiving, organizing data into standardized time-interval intervals with pre-computed severity metrics. This preliminary structuring enables efficient historical queries and analysis while minimizing the volume of stored data by eliminating redundancy and retaining only essential information for trend analysis.
Data Source
AI summary
Methods and systems that use a plurality of machine learning models to both monitor user-generated data entries corresponding to differences in network traffic that may be evidence of a disruption and determine severity levels based on: (i) current and historic differences in average network traffic over the plurality of communication networks; (ii) current and historic user-generated data entries; and (iii) labeled severity levels for historic differences in average network traffic over the plurality of communication networks.