Interaction Data Anomaly Detection for Device Intrusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for detecting device compromise, such as antivirus software and user authentication, often fail to detect sophisticated malware and unauthorized access, lacking an integrated approach to address the multi-faceted nature of modern threats.
Innovation Solution
A system and method for computing device intrusion detection via machine learning that analyzes interaction data, including keystrokes and touch events, to identify schema mismatches and anomalies, providing real-time detection of unauthorized use.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional authentication protocols are used for device security, then device access control is maintained, but detection of sophisticated malware and unauthorized access fails
Solution Approach 1:
The patent replaces traditional authentication protocols (mechanical verification systems) with machine learning-based behavioral analysis. The system uses ML models to analyze interaction data patterns (keystrokes, touch events, accelerometer data) instead of relying on conventional authentication mechanisms, enabling detection of sophisticated threats that traditional systems miss.
Solution Approach 2:
The patent introduces an intermediary machine learning system that sits between the device and threats. This ML-based intermediary analyzes interaction data and schemas to detect anomalies and unauthorized access, acting as a mediator that enhances security without requiring changes to core authentication infrastructure.
2Measurement precision
If comprehensive interaction data analysis is performed for intrusion detection, then detection accuracy improves, but resource consumption increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing interaction data and pre-training ML models with comprehensive datasets before deployment. The system creates schemas for authorized devices in advance and trains models offline, so that during runtime, the actual analysis requires minimal computational resources while maintaining high detection precision.
Solution Approach 2:
The patent applies partial action by selectively analyzing specific interaction data elements (keystrokes, touch events, accelerometer data) based on threat context rather than processing all possible data continuously. The system processes data at appropriate levels of detail, avoiding excessive computation while maintaining detection accuracy.
3Adaptability or versatility
If machine learning models are trained with multiple schemas for authorized devices, then unauthorized access detection capability improves, but training time and computational overhead increase
Solution Approach 1:
The patent segments the training process by creating separate schemas for different authorized devices and training the ML model in modular stages. Each device schema is processed independently, allowing parallel training and reducing overall training time while maintaining the ability to detect unauthorized access across multiple devices.
Data Source
AI summary
Systems, computer program products, and methods are described herein for computing device intrusion detection via machine learning for interaction data analysis. The present disclosure includes receiving interaction event data, associating the interaction event data with the first endpoint device as a first schema, receiving an interaction event data stream and a corresponding endpoint device identifier, determining, by inputting the interaction event data stream and the corresponding endpoint device identifier to a trained machine learning model, an identified endpoint device and a presence of at least one anomaly, and transmitting a notification signal comprising schema mismatch details to the identified endpoint device.